WO2023185714A1 - Computer task processing method and related device therefor - Google Patents

Abstract

Provided in the present application are a computer task processing method and a related device therefor, which are applied to the technical field of computers, and can effectively reduce the time consumed during an equation presolving process, and can also improve the performance of an equation solving process. The method of the present application comprises: acquiring a mixed integer programming equation, which is used for describing a task to be processed, and then performing structure identification on the mixed integer programming equation, so as to obtain information related to the structure of the mixed integer programming equation; subsequently, on the basis of the information, selecting from an operator resource pool at least one operator, which is used for implementing presolving, wherein these selected operators are target operators; and then, implementing presolving for the mixed integer programming equation on the basis of the target operators, and solving the presolved mixed integer programming equation, so as to obtain a corresponding solution, wherein the solution may serve as a processing result of the task.

Description

A computer task processing method and related equipment

This application claims priority to the Chinese patent application filed with the China Patent Office on March 31, 2022, with the application number 202210333041.7 and the invention title "A computer task processing method and related equipment", the entire content of which is incorporated by reference. in this application.

Technical field

The present application relates to the field of computer technology, and in particular, to a computer task processing method and related equipment.

Background technique

Many practical problems faced in many fields can be handled by mixed integer programming. Mixed integer programming usually includes two types of problems: linear programming and nonlinear programming. For example, assuming that a task to be processed is to obtain the factory's production plan, the generated plan not only needs to meet the goals of maximizing factory demand and minimizing the total factory cost, but also needs to meet the hard constraints and soft constraints specified by the actual business scenario. etc., the mathematical model abstracted based on the requirements that the production plan needs to meet can be simplified into the corresponding linear programming equation. Then, by solving the linear programming equation, the production plan of the factory can be obtained.

Since the amount of data involved in the mixed integer programming equation itself is often very large, for example, the size of the variables and constraint matrices in the mixed integer programming equation usually exceeds the order of millions, or even reaches the order of tens of millions, etc., so solutions provided by related technologies can be used The processor preprocesses (presolve) the mixed integer programming equation to reduce the size of the variables and constraint matrices in the equation, thereby obtaining the preprocessed (simplified) mixed integer programming equation, and then performs the preprocessed mixed integer programming equation Solve, which makes solving the equation less difficult.

In the solver, multiple operators for preprocessing are provided, so the solver can execute all operators on the mixed integer programming equation in a certain order to achieve preprocessing of the mixed integer programming equation. However, different mixed integer programming equations may be used to describe the tasks to be processed in different fields. Some mixed integer programming equations may not require the use of all operators to achieve preprocessing. Therefore, existing solvers are in the process of preprocessing. , the factors considered are relatively single, causing the preprocessing process to consume too much time.

Contents of the invention

Embodiments of the present application provide a computer task processing method and related equipment, which can effectively reduce the time consumed in the equation preprocessing process and improve the performance of the equation solving process.

A first aspect of the embodiments of the present application provides a computer task processing method, which method includes:

When a task needs to be processed, the mixed integer programming equation used to describe the task to be processed can be obtained (mixed integer programming refers to that in some planning problems, the elements in the variables can be either integers or non-integers) , mixed integer programming equations usually include two major types of equations, namely linear programming equations and nonlinear programming equations. For example, assume that the task to be processed is to obtain (solve) a production plan to be solved for a certain factory in the supply chain field. The generated plan to be solved not only needs to maximize the demand of the factory and minimize the total cost of the factory. ization and other goals, and also need to meet the hard constraints and soft constraints specified by the actual business scenario, etc. Then, the mathematical model abstracted based on the production plan to be solved and the requirements that the production plan to be solved can be simplified into the corresponding linear programming equation. As a variable in this equation, the The production plan to be solved is usually presented in the form of a vector, so the elements contained in the production plan to be solved can be integers or non-integers. For example, the production plan to be solved contains 10 elements. The first element represents the quantity of raw materials that the factory needs to produce on day 1, the second element represents the quantity of raw materials that the factory needs to produce on day 2,..., The 10th element represents the quantity of raw materials that the factory needs to produce on the 10th day. Therefore, the values of these elements can be either integers (for example, the factory needs to produce 1 ton of raw materials on the 1st day) or non-integers (for example, on the 2nd day The factory needs to produce 0.5 tons of raw materials per day).

After the mixed integer programming equation is obtained, the structure of the mixed integer programming equation can be identified, thereby obtaining information related to the structure of the mixed integer programming equation, which is used to indicate the structure of the mixed integer programming equation. It is worth noting that the operation of structure identification mainly focuses on analyzing the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation, thereby determining the structural composition of the mixed integer programming equation. Therefore, the obtained results are consistent with those of the mixed integer programming equation. The structure-related information is also associated with at least one of the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation.

After obtaining information related to the structure of the mixed integer programming equation, the target operator can be determined from the operator resource pool based on this information. It should be noted that the operator resource pool provides multiple operators for selection. Therefore, based on the information related to the structure of the mixed integer programming equation, the operator resource pool can be selected to implement the prediction of the mixed integer programming equation. At least one operator is processed, and the selected operators are the target operators.

After the target operator is determined, preprocessing (presolve) of the mixed integer programming equation can be implemented based on the target operator, thereby obtaining the preprocessed mixed integer programming equation.

After obtaining the preprocessed mixed integer programming equation, the preprocessed mixed integer programming equation can be solved to obtain the corresponding solution, which can be used as the processing result of the aforementioned task (for example, the solved production plan of a certain factory) .

It can be seen from the above method that after obtaining the mixed integer programming equation used to describe the task to be processed, the target operator can be determined in the operator resource pool based on the information related to the structure of the mixed integer programming equation. This information is related to At least one of the following is relevant: the variables of a mixed integer programming equation and the constraint matrix of a mixed integer programming equation. Then, the mixed integer programming equation can be preprocessed based on the target operator to obtain the preprocessed mixed integer programming equation. Finally, the preprocessed mixed integer programming equation can be solved, and the obtained solution can be used as the processing result of the task. In the foregoing process, the variables and constraint matrices of the mixed integer programming equation can be identified, thereby providing information related to the structure of the mixed integer programming equation. Based on this information, the structure of the mixed integer programming equation can be determined, that is, the mixed integer programming equation belongs to Which type of mixed integer programming equation (used to describe a specific type of task), so as to select a specific target operator for the mixed integer programming equation, thereby realizing the preprocessing of the mixed integer programming equation, which can effectively reduce the cost of the equation preprocessing process time consumed and improves the performance of the equation solving process.

In a possible implementation, since the information related to the structure of the mixed integer programming equation is associated with at least one of the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation, The structure-related information may include at least one of the following: (1) The dimensions of the variables of the mixed integer programming equation, that is, the number of elements contained in the variables of the mixed integer programming equation. (2) The sparsity of the constraint matrix of the mixed integer programming equation, that is, the number of zero-valued elements in the constraint matrix of the mixed integer programming equation. (3) Block information of the constraint matrix of the mixed integer programming equation. The block information of the mixed integer programming equation is used to indicate the submatrix with zero elements contained in the constraint matrix of the mixed integer programming equation. That is, the block information is used In the constraint matrix indicating the mixed integer programming equation, there is a certain region (several regions) in which the elements are all zero. (4) The rows in the constraint matrix of the mixed integer programming equation have a calculation relationship with the variables of the mixed integer programming equation. Since each row in the constraint matrix of the mixed integer programming equation will be multiplied by the variables of the mixed integer programming equation, if the mixed integer programming equation is mixed A row (all elements) of a row in the constraint matrix of an integer programming equation is zero, It means that there is no calculation relationship between this row and the variables of the mixed integer programming equation. If a certain row (any element) in the constraint matrix of the mixed integer programming equation is not zero, it means that there is no calculation relationship between this row and the variables of the mixed integer programming equation. There is a calculation relationship between them.

In a possible implementation, based on information related to the structure of the mixed integer programming equation, determining the target operator in the operator resource pool includes: in the equation resource pool, determining the information related to the structure of the mixed integer programming equation. The preset equation that matches the information; in the operator resource pool, determine the operator used for preprocessing of the preset equation as the target operator. In the aforementioned implementation, the equation resource pool contains multiple preset equations with different structures (which can also be understood as equation templates with different structures set in advance), and information related to the structure of the mixed integer programming equation is used to indicate mixed integer programming. The structural category of the equation, therefore, is to traverse the equation resource pool to find a preset equation that matches the information related to the structure of the mixed integer programming equation. Then, the preset equation and the mixed integer programming equation have a similar structure or The same two equations. After determining the preset equations with similar or identical structures to the mixed integer programming equations, since the preset equations are set in advance, it can be determined which operators in the operator resource pool are used in the preprocessing of the preset equations, and This part of the operators is determined as the operators used to realize the preprocessing of the mixed integer programming equation, that is, the target operator. It can be seen that the embodiment of the present application provides an equation resource pool. After determining which type of preset mixed integer programming equation the mixed integer programming equation belongs to, that is, a certain preset equation in the equation resource pool (which can also be called an equation resource pool) (an equation template in ), the operator used for preprocessing of the preset equation can be used as the operator used for preprocessing of the mixed integer programming equation to reduce the impact of human intervention on preprocessing, thereby Preprocessing effects of optimization equations.

In a possible implementation, in the preprocessing of mixed integer programming equations, the number of target operators is n, and the number of iterations of the n target operators is k times, and both n and k are greater than or equal to 1. Integer of Before the i-th operator, you can first detect whether the sum of the indicators from the i-th target operator in the first iteration to the i-th target operator in the j-1th iteration is less than the first threshold (th A threshold is usually determined based on the number of iterations k, and generally increases with the increase of the number of iterations k, that is, there is a positive correlation between the two). Among them, i=1,...,n, j=1,...,k. If the sum of the index of the i-th target operator in the 1st iteration to the index of the i-th target operator in the j-1th iteration is less than the first threshold, it means that the i-th target operator needs to continue to be executed thereafter. , then further detect whether the index of the i-th target operator in the j-1th iteration is less than the second threshold (the second threshold is usually determined based on the number of iterations k, and generally increases with the increase of the number of iterations k, That is, there is a positive correlation between the two). If the sum of the index of the i-th target operator in the 1st iteration to the index of the i-th target operator in the j-1th iteration is greater than or equal to the first threshold, it means that there is no need to continue executing the i-th operation in the future. target operator, then in the j-th iteration to the k-th iteration, the i-th target operator is skipped. If the index of the i-th target operator in the j-1th iteration is less than the second threshold, it means that the i-th operator can currently be executed, then the i-th target operator is executed on the updated mixed integer programming equation, and Generate the index of the i-th target operator in the j-th iteration. It is worth noting that the index of the i-th target operator in the j-th iteration can be generated in the following way: first, determine the rows deleted by the i-th target operator in the 1st iteration to the j-th iteration. in the updated mixed integer programming equation. Then, calculation is performed based on this proportion to obtain the index of the i-th target operator in the j-th iteration. If the index of the i-th target operator in the j-1th iteration is greater than or equal to the second threshold, it means that the i-th operator cannot be executed currently, then jump to the i+1-th target operator, and perform the operation of the i-th target operator. i+1 target operators execute the above steps until the jth iteration of n target operators is completed. Continue to execute the j+1th iteration of n target operators until the kth iteration of n target operators is completed, and obtain the preprocessed mixed integer programming equation. In the aforementioned implementation, Implement the preprocessing process with a certain iteration strategy, that is, during the iteration process, when the role of an operator becomes smaller, it can be considered that the benefit of continuing to execute the operator in subsequent iterations is limited, and the execution of the operator should be stopped in time . Moreover, a corresponding stopping mechanism is also designed for the preprocessing cycle. In this way, not only the mixed integer programming equation can be effectively simplified, but also unnecessary preprocessing processes can be reduced, thereby reducing the time overhead of preprocessing.

In a possible implementation, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the number of iterations of the n target operators is the same. Furthermore, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the execution order of the n target operators is the same. In the aforementioned implementation method, after determining which type of preset mixed integer programming equation the mixed integer programming equation belongs to, that is, a certain preset equation in the equation resource pool, the algorithm used for preprocessing of the preset equation can be The number of sub-iterations and the order of operator execution are used as the number of operator iterations and the order of operator execution used for preprocessing of the mixed integer programming equation to reduce the impact of human intervention on preprocessing and thereby optimize the preprocessing effect of the equation.

In one possible implementation, the index of the i-th target operator in the j-th iteration is determined based on the proportion of rows deleted by the i-th target operator in the mixed integer programming equation; further, as mentioned above The first threshold and the aforementioned second threshold are determined based on the number of iterations of n target operators.

In one possible implementation, the mixed integer programming equation is preprocessed based on the target operator, and before the preprocessed mixed integer programming equation is obtained, the method further includes: merging rows with a multiple relationship in the constraint matrix, or , set the multiple columns in the constraint matrix to the same column to obtain the constraint matrix after the redundancy is removed; update the mixed integer programming equation based on the constraint matrix after the redundancy is removed, and obtain the updated mixed integer programming equation; Preprocessing the mixed integer programming equation based on the objective operator to obtain the preprocessed mixed integer programming equation includes: preprocessing the updated mixed integer programming equation based on the objective operator to obtain the preprocessed mixed integer programming equation. In the aforementioned implementation method, in order to achieve better preprocessing, the mixed integer programming equation can be deredundant in advance. This operation can be any of the following three situations: (1) In the constraints of the mixed integer programming equation In the matrix, the rows with a multiple relationship can be determined in a certain way, and these rows can be merged to obtain the constraint matrix after the merged rows. In the constraint matrix after merging rows, the columns with multiple relationships can be determined in the same way, and these columns are set to the same columns to obtain the constraint matrix after redundancy is removed. (2) In the constraint matrix of the mixed integer programming equation, the rows with multiple relationships can be determined in a certain way, and these rows can be merged to obtain the constraint matrix after removing redundancy. (3) In the constraint matrix of the mixed integer programming equation, the columns with a multiple relationship can be determined in a certain way, and these columns are set to the same column to obtain the constraint matrix after redundancy is removed. After obtaining the constraint matrix after redundancy removal based on the aforementioned redundancy removal operation, the original mixed integer programming equation can be updated with the redundancy removal constraint matrix to obtain the updated mixed integer programming equation. In this way, the preprocessing object is transformed from the original mixed integer programming equation to the updated mixed integer programming equation, which can optimize the preprocessing effect.

The second aspect of the embodiment of the present application provides a method for obtaining a production plan. The method includes: obtaining a mixed integer programming equation, which is used to describe the production plan to be solved; based on the structure related to the mixed integer programming equation. information, determine the target operator in the operator resource pool, and the target operator is part of the operators in the operator resource pool; preprocess the mixed integer programming equation based on the target operator to obtain the preprocessed mixed integer programming equation; The preprocessed mixed integer programming equation is solved to obtain the solved production plan.

It can be seen from the above method that after obtaining the mixed integer programming equation used to describe the production plan to be solved, the target operator can be determined in the operator resource pool based on the information related to the structure of the mixed integer programming equation. This information Related to at least one of the following: variables of a mixed integer programming equation and constraint matrices of a mixed integer programming equation. Then, based on the goals The operator preprocesses the mixed integer programming equation and obtains the preprocessed mixed integer programming equation. Finally, the preprocessed mixed integer programming equation can be solved to obtain the solved production plan. In the foregoing process, the variables and constraint matrices of the mixed integer programming equation can be identified, thereby providing information related to the structure of the mixed integer programming equation. Based on this information, the structure of the mixed integer programming equation can be determined, that is, the mixed integer programming equation belongs to Which type of mixed integer programming equation (used to solve the production plan of factories and enterprises in the field of supply chain), so as to select a specific target operator for this mixed integer programming equation, so as to achieve the preprocessing of this mixed integer programming equation, can be effective Reduce the time consumed by the equation preprocessing process and improve the performance of the equation solving process.

In a possible implementation, the variable of the mixed integer programming equation is the production plan to be solved, and the constraint matrix of the mixed integer programming equation includes at least one of the following: material constraints, productivity constraints, and time that the production plan to be solved needs to satisfy. Processing constraints such as constraints or traffic constraints. Among them, material constraints can include the upper limit of materials that can be used in production and material ratios, etc. Productivity constraints can include the upper limit of the number of production workers required during production, the upper limit of work efficiency, etc., time Constraints may include the delivery date of the product, etc. Transportation constraints may include the upper limit of transportation costs required during production and the upper limit of the number of transportation vehicles required during production, etc.

In a possible implementation, the information related to the structure of the mixed integer programming equation includes at least one of the following: the dimensions of the variables, the sparsity of the constraint matrix, the block information of the constraint matrix, or the existence of a computational relationship with the variables in the constraint matrix. rows, where the blocking information is used to indicate that the constraint matrix contains submatrices with zero elements.

In a possible implementation, based on information related to the structure of the mixed integer programming equation, determining the target operator in the operator resource pool includes: in the equation resource pool, determining the information related to the structure of the mixed integer programming equation. The preset equation that matches the information; in the operator resource pool, the operator corresponding to the preprocessing of the preset equation is determined as the target operator.

In a possible implementation, in the preprocessing of mixed integer programming equations, the number of target operators is n, and the number of iterations of the n target operators is k times, and both n and k are greater than or equal to 1. Integer of If the sum of indicators from the indicators in the iteration to the i-th target operator in the j-1 iteration is greater than or equal to the first threshold, then the i-th target is skipped in the j-th iteration to the k-th iteration. Operator, i=1,...,n,j=1,...,k; or if the sum is less than the first threshold, detect whether the index of the i-th target operator in the j-1th iteration is less than the second threshold ; Or, if the index of the i-th objective operator in the j-1th iteration is less than the second threshold, execute the i-th objective operator on the mixed integer programming equation and generate the i-th objective operator in the j-th iteration. index in iterations; or, if the index of the i-th target operator in the j-1th iteration is greater than or equal to the second threshold, then perform the above steps for the i+1th target operator until n The jth iteration of the target operator; execute the j+1 iteration of the n target operators until the kth iteration of the n target operators is completed, and obtain the preprocessed mixed integer programming equation.

In a possible implementation, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the number of iterations of the n target operators is the same.

In a possible implementation, the index of the i-th target operator in the j-th iteration is determined based on the proportion of rows deleted by the i-th target operator in the mixed integer programming equation.

In a possible implementation, the first threshold and the second threshold are determined based on the number of iterations of n target operators.

In one possible implementation, the mixed integer programming equation is preprocessed based on the target operator, and before the preprocessed mixed integer programming equation is obtained, the method further includes: merging rows with a multiple relationship in the constraint matrix, or , set the multiple columns in the constraint matrix to the same column to obtain the constraint matrix after the redundancy is removed; based on the constraints after the redundancy is removed The matrix updates the mixed integer programming equation to obtain the updated mixed integer programming equation; the mixed integer programming equation is preprocessed based on the target operator, and the preprocessed mixed integer programming equation is obtained, including: based on the target operator, the updated mixed integer programming equation is obtained The mixed integer programming equation is preprocessed to obtain the preprocessed mixed integer programming equation.

The third aspect of the embodiment of the present application provides a power grid dispatching method. The method includes: obtaining a mixed integer programming equation, which is used to describe the power consumption dispatch plan to be solved; based on the structure of the mixed integer programming equation Relevant information, determine the target operator in the operator resource pool, and the target operator is part of the operators in the operator resource pool; preprocess the mixed integer programming equation based on the target operator to obtain the preprocessed mixed integer programming Equation; solve the preprocessed mixed integer programming equation to obtain the solved power consumption dispatch plan.

The fourth aspect of the embodiment of the present application provides a central site selection method, which method includes: obtaining a mixed integer programming equation, which is used to describe the address selection plan to be solved; based on the structural correlation with the mixed integer programming equation information, determine the target operator in the operator resource pool, and the target operator is some of the operators in the operator resource pool; preprocess the mixed integer programming equation based on the target operator to obtain the preprocessed mixed integer programming equation ; Solve the preprocessed mixed integer programming equation to obtain the solved address selection plan.

The fifth aspect of the embodiment of the present application provides a computer task processing device. The device includes: an acquisition module for acquiring a mixed integer programming equation, which is used to describe the task to be processed; and a determination module for based on Information related to the structure of the mixed integer programming equation, determine the target operator in the operator resource pool, and the target operator is part of the operators in the operator resource pool; the preprocessing module is used to perform mixed integer planning based on the target operator The equation is preprocessed to obtain the preprocessed mixed integer programming equation; the solving module is used to solve the preprocessed mixed integer programming equation, and the obtained solution is used as the processing result of the task.

It can be seen from the above device that after obtaining the mixed integer programming equation used to describe the task to be processed, the target operator can be determined in the operator resource pool based on the information related to the structure of the mixed integer programming equation. This information is related to At least one of the following is relevant: the variables of a mixed integer programming equation and the constraint matrix of a mixed integer programming equation. Then, the mixed integer programming equation can be preprocessed based on the target operator to obtain the preprocessed mixed integer programming equation. Finally, the preprocessed mixed integer programming equation can be solved, and the obtained solution can be used as the processing result of the task. In the foregoing process, the variables and constraint matrices of the mixed integer programming equation can be identified, thereby providing information related to the structure of the mixed integer programming equation. Based on this information, the structure of the mixed integer programming equation can be determined, that is, the mixed integer programming equation belongs to Which type of mixed integer programming equation (used to describe a specific type of task), so as to select a specific target operator for the mixed integer programming equation, thereby realizing the preprocessing of the mixed integer programming equation, which can effectively reduce the cost of the equation preprocessing process time consumed and improves the performance of the equation solving process.

In a possible implementation, the information related to the structure of the mixed integer programming equation includes at least one of the following: the dimensions of the variables, the sparsity of the constraint matrix, the block information of the constraint matrix, and the existence of a calculation relationship with the variables in the constraint matrix. rows, where the blocking information is used to indicate that the constraint matrix contains submatrices with zero elements.

In a possible implementation, the determination module is used to: in the equation resource pool, determine the preset equation that matches the information related to the structure of the mixed integer programming equation; in the operator resource pool, determine the preset equation for The operator used in the preprocessing of the equation is determined as the target operator.

In a possible implementation, in the preprocessing of mixed integer programming equations, the number of target operators is n, and the number of iterations of the n target operators is k times, and both n and k are greater than or equal to 1. Integer of If the sum of the indicators in the j-1 iteration is greater than or equal to the first threshold, then in the j-th iteration to the k-th iteration, the i-th target operator is skipped, i=1,...,n, j=1,...,k; or, if the sum of the index of the i-th target operator in the 1st iteration to the index of the i-th target operator in the j-1th iteration is less than the first threshold, then Detect whether the index of the i-th target operator in the j-1th iteration is less than the second threshold; or, if the index of the i-th target operator in the j-1th iteration is less than the second threshold, then The integer programming equation executes the i-th objective operator and generates the indicator of the i-th objective operator in the j-th iteration; or, if the indicator of the i-th objective operator in the j-1 iteration is greater than or equal to second threshold, then perform the above steps for the i+1th target operator until the jth iteration of the n target operators is completed; perform the j+1th iteration of the n target operators until the n targets are completed. After the kth iteration of the operator, the preprocessed mixed integer programming equation is obtained.

In a possible implementation, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the number of iterations of the n target operators is the same.

In a possible implementation, the index of the i-th target operator in the j-th iteration is determined based on the proportion of rows deleted by the i-th target operator in the mixed integer programming equation.

In a possible implementation, the first threshold and the second threshold are determined based on the number of iterations of n target operators.

In a possible implementation, the device further includes: a redundancy removal module, used to merge the rows with a multiple relationship in the constraint matrix, or set the columns with a multiple relationship in the constraint matrix to the same column, to obtain The constraint matrix after the redundancy is removed; the update module is used to update the mixed integer programming equation based on the redundant constraint matrix to obtain the updated mixed integer programming equation; the preprocessing module is used to update the mixture based on the target operator The final mixed integer programming equation is preprocessed to obtain the preprocessed mixed integer programming equation.

The sixth aspect of the embodiment of the present application provides a production plan acquisition device. The device includes: an acquisition module for acquiring a mixed integer programming equation. The mixed integer programming equation is used to describe the production plan to be solved in the field of supply chain; determine The module is used to determine the target operator in the operator resource pool based on the information related to the structure of the mixed integer programming equation. The target operator is part of the operators in the operator resource pool; the preprocessing module is used to determine the target operator based on the target operator resource pool. The module preprocesses the mixed integer programming equation to obtain the preprocessed mixed integer programming equation; the solving module is used to solve the preprocessed mixed integer programming equation to obtain the solved production plan.

It can be seen from the above device that after obtaining the mixed integer programming equation used to describe the production plan to be solved, the target operator can be determined in the operator resource pool based on the information related to the structure of the mixed integer programming equation. This information Related to at least one of the following: variables of a mixed integer programming equation and constraint matrices of a mixed integer programming equation. Then, the mixed integer programming equation can be preprocessed based on the target operator to obtain the preprocessed mixed integer programming equation. Finally, the preprocessed mixed integer programming equation can be solved to obtain the solved production plan. In the foregoing process, the variables and constraint matrices of the mixed integer programming equation can be identified, thereby providing information related to the structure of the mixed integer programming equation. Based on this information, the structure of the mixed integer programming equation can be determined, that is, the mixed integer programming equation belongs to Which type of mixed integer programming equation (used to solve the production plan of factories and enterprises in the field of supply chain), so as to select a specific target operator for this mixed integer programming equation, so as to achieve the preprocessing of this mixed integer programming equation, can be effective Reduce the time consumed by the equation preprocessing process and improve the performance of the equation solving process.

In a possible implementation, the variable of the mixed integer programming equation is the production plan to be solved, and the constraint matrix of the mixed integer programming equation includes at least one of the following: material constraints, productivity constraints, and time that the production plan to be solved needs to satisfy. restraints or traffic constraints.

In a possible implementation, the information related to the structure of the mixed integer programming equation includes at least one of the following: The dimension of the variable, the sparsity of the constraint matrix, the blocking information of the constraint matrix, or the rows in the constraint matrix that have a computational relationship with the variable, where the blocking information is used to indicate the submatrix whose elements are zero contained in the constraint matrix.

In a possible implementation, the determination module is used to: in the equation resource pool, determine the preset equation that matches the information related to the structure of the mixed integer programming equation; in the operator resource pool, determine the preset equation for The operator corresponding to the preprocessing of the equation is determined as the target operator.

In a possible implementation, in the preprocessing of mixed integer programming equations, the number of target operators is n, and the number of iterations of the n target operators is k times, and both n and k are greater than or equal to 1. is an integer, preprocessing module, used for: in the j-th iteration of n target operators, if the index of the i-th target operator in the 1st iteration to the i-th target operator is in the j-1 If the sum of indicators in the iteration is greater than or equal to the first threshold, then in the j-th iteration to the k-th iteration, the i-th target operator is skipped, i=1,...,n,j=1,..., k; or if the sum is less than the first threshold, then detect whether the index of the i-th target operator in the j-1th iteration is less than the second threshold; or, if the i-th target operator in the j-1th iteration If the index in is less than the second threshold, execute the i-th target operator on the mixed integer programming equation and generate the index of the i-th target operator in the j-th iteration; or, if the i-th target operator is in the j-th iteration If the indicator in j-1 iterations is greater than or equal to the second threshold, perform the above steps for the i+1th target operator until the jth iteration of n target operators is completed; execute the nth iteration of n target operators. j+1 iterations until the k-th iteration of n target operators is completed, and the preprocessed mixed integer programming equation is obtained.

In a possible implementation, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the number of iterations of the n target operators is the same.

In a possible implementation, the index of the i-th target operator in the j-th iteration is determined based on the proportion of rows deleted by the i-th target operator in the mixed integer programming equation.

In a possible implementation, the first threshold and the second threshold are determined based on the number of iterations of n target operators.

In a possible implementation, the device further includes: a redundancy removal module, used to merge the rows with a multiple relationship in the constraint matrix, or set the columns with a multiple relationship in the constraint matrix to the same column, to obtain The constraint matrix after the redundancy is removed; the update module is used to update the mixed integer programming equation based on the redundant constraint matrix to obtain the updated mixed integer programming equation; the preprocessing module is used to update the mixture based on the target operator The final mixed integer programming equation is preprocessed to obtain the preprocessed mixed integer programming equation.

The seventh aspect of the embodiment of the present application provides a power grid dispatching device. The device includes: an acquisition module, used to obtain a mixed integer programming equation, which is used to describe the power consumption dispatch plan to be solved; a determination module, It is used to determine the target operator in the operator resource pool based on the information related to the structure of the mixed integer programming equation. The target operator is a part of the operators in the operator resource pool; the preprocessing module is used to determine the target operator based on the target operator. The mixed integer programming equation is preprocessed to obtain the preprocessed mixed integer programming equation; the solving module is used to solve the preprocessed mixed integer programming equation to obtain the solved power consumption dispatch plan.

The eighth aspect of the embodiment of the present application provides a center site selection method. The device includes: an acquisition module, used to obtain a mixed integer programming equation, which is used to describe the address selection plan to be solved; a determination module, using Based on the information related to the structure of the mixed integer programming equation, the target operator is determined in the operator resource pool, and the target operator is part of the operators in the operator resource pool; the preprocessing module is used to pair the mixture based on the target operator The integer programming equation is preprocessed to obtain the preprocessed mixed integer programming equation; the solving module is used to solve the preprocessed mixed integer programming equation to obtain the solved address selection plan.

A ninth aspect of the embodiment of the present application provides a computer task processing device. The device includes: a processor, a memory, a bus, and an input and output device. The processor is connected to the memory and the input and output devices. The bus is connected to the processor, the memory, and the input and output devices respectively. The input and output devices are connected; the input and output devices are used to obtain mixed integer programming equations, and the mixed integer programming equations are used to describe the tasks to be processed; an operator resource pool is stored in the memory; the processor is used to: obtain the operator resource pool from the memory , and based on the information related to the structure of the mixed integer programming equation, determine the target operator in the operator resource pool, and the target operator is some of the operators in the operator resource pool; predict the mixed integer planning equation based on the target operator Process to obtain the preprocessed mixed integer programming equation; solve the preprocessed mixed integer programming equation, and the obtained solution is used as the processing result of the task.

In a possible implementation, the information related to the structure of the mixed integer programming equation includes at least one of the following: dimensions of variables of the mixed integer programming equation, sparsity of the constraint matrix of the mixed integer programming equation, and block partitioning of the constraint matrix. The rows in the information or constraint matrix that have a computational relationship with the variables, where the blocking information is used to indicate that the constraint matrix contains a submatrix with zero elements.

In a possible implementation, the processor is configured to: in the equation resource pool, determine a preset equation that matches the information related to the structure of the mixed integer programming equation; in the operator resource pool, determine the preset equation for The operator corresponding to the preprocessing of the equation is determined as the target operator.

In a possible implementation, in the preprocessing of mixed integer programming equations, the number of target operators is n, and the number of iterations of the n target operators is k times, and both n and k are greater than or equal to 1. is an integer, processor, used for: in the j-th iteration of n target operators, if the index of the i-th target operator in the 1st iteration to the i-th target operator in the j-1th iteration If the sum of indicators in the iteration is greater than or equal to the first threshold, then in the j-th iteration to the k-th iteration, the i-th target operator is skipped, i=1,...,n, j=1,...,k ; Or if the sum is less than the first threshold, then detect whether the index of the i-th target operator in the j-1th iteration is less than the second threshold; or, if the i-th target operator in the j-1th iteration The index of is less than the second threshold, then execute the i-th target operator on the mixed integer programming equation and generate the index of the i-th target operator in the j-th iteration; or, if the i-th target operator is in the j-th iteration If the indicator in the -1 iteration is greater than or equal to the second threshold, perform the above steps for the i+1th target operator until the jth iteration of n target operators is completed; execute the jth iteration of n target operators +1 iterations until the k-th iteration of n target operators is completed, and the preprocessed mixed integer programming equation is obtained.

In a possible implementation, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the number of iterations of the n target operators is the same.

In a possible implementation, the index of the i-th target operator in the j-th iteration is determined based on the proportion of rows deleted by the i-th target operator in the mixed integer programming equation.

In a possible implementation, the first threshold and the second threshold are determined based on the number of iterations of n target operators.

In a possible implementation, the processor is also used to: merge rows that are in a multiple relationship in the constraint matrix, or set columns that are in a multiple relationship in the constraint matrix to the same column to obtain the deredundant Constraint matrix; update the mixed integer programming equation based on the redundantly removed constraint matrix to obtain the updated mixed integer programming equation; processor, used for: preprocessing the updated mixed integer programming equation based on the target operator, Obtain the preprocessed mixed integer programming equation.

The tenth aspect of the embodiment of the present application provides a production plan acquisition device. The device includes: a processor, a memory, a bus, and an input and output device. The processor is connected to the memory and the input and output devices. The bus is connected to the processor, the memory, and the input and output devices respectively. The input and output devices are connected; the input and output devices are used to obtain mixed integer programming equations, which are used to describe the production plan to be solved in the field of supply chain; an operator resource pool is stored in the memory; the processor is used to:

Obtain the operator resource pool from the memory, and determine the target operator in the operator resource pool based on the information related to the structure of the mixed integer programming equation. The target operator is some of the operators in the operator resource pool; based on the target The operator preprocesses the mixed integer programming equation to obtain the preprocessed mixed integer programming equation; it solves the preprocessed mixed integer programming equation to obtain the solved production plan.

In a possible implementation, the variable of the mixed integer programming equation is the production plan to be solved, and the constraint matrix of the mixed integer programming equation includes at least one of the following: material constraints, productivity constraints, and time that the production plan to be solved needs to satisfy. restraints or traffic constraints.

In a possible implementation, the information related to the structure of the mixed integer programming equation includes at least one of the following: dimensions of variables of the mixed integer programming equation, sparsity of the constraint matrix of the mixed integer programming equation, and block partitioning of the constraint matrix. The rows in the information or constraint matrix that have a computational relationship with the variables, where the blocking information is used to indicate that the constraint matrix contains a submatrix with zero elements.

In a possible implementation, the processor is configured to: in the equation resource pool, determine a preset equation that matches the information related to the structure of the mixed integer programming equation; in the operator resource pool, determine the preset equation for The operator corresponding to the preprocessing of the equation is determined as the target operator.

In a possible implementation, in the preprocessing of mixed integer programming equations, the number of target operators is n, and the number of iterations of the n target operators is k times, and both n and k are greater than or equal to 1. is an integer, processor, used for: in the j-th iteration of n target operators, if the index of the i-th target operator in the 1st iteration to the i-th target operator in the j-1th iteration If the sum of indicators in the iteration is greater than or equal to the first threshold, then in the j-th iteration to the k-th iteration, the i-th target operator is skipped, i=1,...,n, j=1,...,k ; Or if the sum is less than the first threshold, then detect whether the index of the i-th target operator in the j-1th iteration is less than the second threshold; or, if the i-th target operator in the j-1th iteration The index of is less than the second threshold, then execute the i-th target operator on the mixed integer programming equation and generate the index of the i-th target operator in the j-th iteration; or, if the i-th target operator is in the j-th iteration If the indicator in the -1 iteration is greater than or equal to the second threshold, perform the above steps for the i+1th target operator until the jth iteration of n target operators is completed; execute the jth iteration of n target operators +1 iterations until the k-th iteration of n target operators is completed, and the preprocessed mixed integer programming equation is obtained.

In a possible implementation, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the number of iterations of the n target operators is the same.

In a possible implementation, the index of the i-th target operator in the j-th iteration is determined based on the proportion of rows deleted by the i-th target operator in the mixed integer programming equation.

In a possible implementation, the first threshold and the second threshold are determined based on the number of iterations of n target operators.

In a possible implementation, the processor is also used to: merge rows that are in a multiple relationship in the constraint matrix, or set columns that are in a multiple relationship in the constraint matrix to the same column to obtain the deredundant Constraint matrix; update the mixed integer programming equation based on the redundantly removed constraint matrix to obtain the updated mixed integer programming equation; processor, used for: preprocessing the updated mixed integer programming equation based on the target operator, Obtain the preprocessed mixed integer programming equation.

The eleventh aspect of the embodiment of the present application provides a power grid dispatching device. The device includes: a processor, a memory, a bus, and an input and output device. The processor is connected to the memory and the input and output devices. The bus is connected to the processor, the memory, and the input and output devices respectively. The input and output devices are connected; the input and output devices are used to obtain mixed integer programming equations, and the mixed integer programming equations are used to describe the power consumption scheduling plan to be solved in the power field; an operator resource pool is stored in the memory; the processor is used to: obtain from the memory Obtain the operator resource pool in the operator resource pool, and based on the information related to the structure of the mixed integer programming equation, in the operator resource pool Determine the target operator, which is part of the operator in the operator resource pool; preprocess the mixed integer programming equation based on the target operator to obtain the preprocessed mixed integer programming equation; perform the preprocessed mixed integer The planning equation is solved to obtain the solved power consumption schedule.

A twelfth aspect of the embodiment of the present application provides a central address selection device. The device includes: a processor, a memory, a bus, and an input and output device. The processor is connected to the memory and the input and output devices, and the bus is connected to the processor and the memory respectively. It is connected to the input and output devices; the input and output devices are used to obtain mixed integer programming equations, and the mixed integer programming equations are used to describe the address selection plan to be solved in the field of cloud technology; an operator resource pool is stored in the memory; the processor is used to:

Obtain the operator resource pool from the memory, and determine the target operator in the operator resource pool based on the information related to the structure of the mixed integer programming equation. The target operator is some of the operators in the operator resource pool; based on the target The operator preprocesses the mixed integer programming equation to obtain the preprocessed mixed integer programming equation; it solves the preprocessed mixed integer programming equation to obtain the solved address selection plan.

A thirteenth aspect of the embodiments of the present application provides a computer storage medium. The computer storage medium stores a computer program. When the program is executed by a computer, the computer implements the first aspect and any possible implementation manner of the first aspect. , the method described in the second aspect or any possible implementation of the second aspect.

A fourteenth aspect of the embodiments of the present application provides a computer program product. The computer program product stores instructions. When the instructions are executed by a computer, the computer implements the first aspect and any possible implementation manner of the first aspect. , the method described in the second aspect or any possible implementation of the second aspect.

In the embodiment of the present application, after obtaining the mixed integer programming equation used to describe the task to be processed, the target operator can be determined in the operator resource pool based on information related to the structure of the mixed integer programming equation. This information is related to the following At least one item is relevant: the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation. Then, the mixed integer programming equation can be preprocessed based on the target operator to obtain the preprocessed mixed integer programming equation. Finally, the preprocessed mixed integer programming equation can be solved, and the obtained solution can be used as the processing result of the task. In the foregoing process, the variables and constraint matrices of the mixed integer programming equation can be identified, thereby providing information related to the structure of the mixed integer programming equation. Based on this information, the structure of the mixed integer programming equation can be determined, that is, the mixed integer programming equation belongs to Which type of mixed integer programming equation (used to describe a specific type of task), so as to select a specific target operator for the mixed integer programming equation, thereby realizing the preprocessing of the mixed integer programming equation, which can effectively reduce the cost of the equation preprocessing process time consumed and improves the performance of the equation solving process.

Description of drawings

Figure 1 is a schematic structural diagram of a computer task processing system provided by an embodiment of the present application;

Figure 2 is another structural schematic diagram of a computer task processing system provided by an embodiment of the present application;

Figure 3a shows an exemplary structural diagram of the terminal device 300;

Figure 3b is an architectural schematic diagram of a solver provided by an embodiment of the present application;

Figure 4 is a schematic flowchart of a computer task processing method provided by an embodiment of the present application;

Figure 5 is a schematic flow chart of preprocessing provided by the embodiment of the present application;

Figure 6 is a schematic diagram of an application example of the computer task processing method provided by the embodiment of the present application;

Figure 7 is a schematic diagram of the material assembly relationship in the production process provided by the embodiment of the present application;

Figure 8 is a schematic diagram of the comparison results provided by the embodiment of the present application;

Figure 9 is a schematic structural diagram of a computer task processing device provided by an embodiment of the present application;

Figure 10 is a schematic structural diagram of a production plan acquisition device provided by an embodiment of the present application.

Detailed ways

Embodiments of the present application provide a computer task processing method and related equipment, which can effectively reduce the time consumed in the equation preprocessing process and improve the performance of the equation solving process.

The terms "first", "second", etc. in the description and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the terms so used are interchangeable under appropriate circumstances, and are merely a way of distinguishing objects with the same attributes in describing the embodiments of the present application. Furthermore, the terms "include" and "having" and any variations thereof, are intended to cover non-exclusive inclusion, such that a process, method, system, product or apparatus comprising a series of elements need not be limited to those elements, but may include any other elements specifically listed or inherent to such processes, methods, products or equipment.

Many practical problems faced in many fields can be handled by mixed integer programming. Mixed integer programming usually includes two types of problems: linear programming and nonlinear programming. For example, assuming that a task to be processed is to obtain the factory's production plan, the generated plan not only needs to meet the goals of maximizing factory demand and minimizing the total factory cost, but also needs to meet the hard constraints and soft constraints specified by the actual business scenario. etc., the mathematical model abstracted based on the requirements that the production plan needs to meet can be simplified into the corresponding linear programming equation. Then, by solving the linear programming equation, the production plan of the factory can be obtained.

Since the amount of data involved in the mixed integer programming equation itself is often very large, for example, the size of the variables and constraint matrices in the mixed integer programming equation (that is, the number of elements in the variable and constraint matrices) usually exceeds the order of millions, or even reaches the order of tens of millions. etc. Therefore, the solver provided by the relevant technology can be used to preprocess (presolve) the mixed integer programming equation to reduce the size of the variables and constraint matrices in the equation, thereby obtaining the preprocessed (simplified) mixed integer programming equation. , and then solve the preprocessed mixed integer programming equation to obtain the corresponding solution. Based on this solution method, the difficulty of solving mixed integer programming equations can be effectively reduced.

In the solver, multiple operators for preprocessing are provided, so the solver can execute all operators on the mixed integer programming equation in a certain order to achieve preprocessing of the mixed integer programming equation. However, different mixed integer programming equations may be used to describe tasks to be processed in different fields. Some mixed integer programming equations may not require the use of all operators to implement preprocessing. More operators are required, and some mixed integer programming equations require fewer operators when implementing preprocessing. Therefore, existing solvers consider only a single factor when preprocessing mixed integer programming equations. This causes the preprocessing process to consume too much time.

Furthermore, when the solver executes multiple operators on the mixed integer programming equation, it is usually executed multiple times in a loop, and the number of iterations of the operators is usually set artificially. However, artificial settings often rely on expert experience, and different mixed integer programming equations may be used to describe tasks to be processed in different fields. For different mixed integer programming equations, artificially setting the number of operator iterations is often not enough. Accurate, resulting in insufficient preprocessing effect of the equation.

In order to solve the above problems, embodiments of the present application provide a computer task processing method (which can also be considered to provide a new solver). This method makes certain improvements on the solvers of related technologies, so that it can Mixed integer programming equations (including linear programming equations, nonlinear programming equations, etc.) used to describe tasks in different fields are preprocessed and solved with high quality to obtain the processing results of the tasks. The methods provided by the embodiments of this application can be applied in a variety of scenarios, The application scenarios of this method will be introduced below:

Figure 1 is a schematic structural diagram of a computer task processing system provided by an embodiment of the present application. The computer task processing system includes user equipment and data processing equipment. Among them, user equipment includes smart terminals such as mobile phones, personal computers, or information processing centers. The user equipment is the initiator of the data sequence processing. As the initiator of the data sequence processing request, the user usually initiates the request through the user equipment.

The above-mentioned data processing equipment may be a cloud server, a network server, an application server, a management server, and other equipment or servers with data processing functions. The data processing device receives the task processing request from the smart terminal through the interactive interface, and then abstracts the task into a mixed data planning equation through the memory for storing data and the processor for data processing, and performs task processing in the form of equation preprocessing and solving. . The memory in the data processing device can be a general term, including local storage and a database that stores historical data. The database can be on the data processing device or on other network servers.

In the computer task processing system shown in Figure 1, the user device can receive instructions from the user. For example, the user device can obtain a task input/selected by the user, and then initiate a request to the data processing device (usually including various parameters of the task, For example, variables of the task and various constraints that the task needs to satisfy, etc.), so that the data processing device executes a task processing application for the task obtained by the user device, thereby obtaining a processing result for the task. For example, the user device can obtain a task input by the user and related parameters of the task, and then initiate a task processing request to the data processing device, so that the data processing device abstracts the task into a mixed integer equation and performs the processing on the mixed integer equation. Preprocessing, and then solving the preprocessed mixed integer equation to obtain the corresponding solution, which can be used as the processing result of the task.

In Figure 1, the data processing device can execute the computer task processing method or the production plan obtaining method according to the embodiment of the present application.

Figure 2 is another structural schematic diagram of a computer task processing system provided by an embodiment of the present application. In Figure 2, the user equipment directly serves as a data processing equipment. The user equipment can directly obtain input from the user and directly use the hardware of the user equipment itself. The specific process is similar to Figure 1. Please refer to the above description and will not be repeated here.

In the computer task processing system shown in Figure 2, the user equipment can receive instructions from the user. For example, the user equipment can obtain a task selected by the user in the user equipment, and then the user equipment itself executes a task processing application for the task. (For example, text summary generation, etc.), that is, abstract the task into a mixed integer equation, preprocess the mixed integer equation, and then solve the preprocessed mixed integer equation to obtain the corresponding solution, which can be used as the task processing results.

In Figure 2, the user equipment itself can execute the computer task processing method or the production plan obtaining method according to the embodiment of the present application.

Further, the user equipment in Figures 1 and 2 can usually be a terminal equipment used by the user, such as a mobile phone, a notebook, a personal computer, etc. In order to facilitate understanding of the terminal device used by the user, the terminal device will be further introduced below in conjunction with Figure 3a (Figure 3a shows an exemplary structural schematic diagram of the terminal device 300). As shown in Figure 3a, the terminal device 300 includes: an application processor 301, a microcontroller unit (MCU) 303, a memory 305, a modem (modem) 307, a radio frequency (radio frequency, RF) module 309, and wireless fidelity. (Wireless-Fidelity, Wi-Fi for short) module 311, Bluetooth module 313, sensor 314, positioning module 350, input/output (input/output, I/O) device 335 and other components. These components communicate via one or more communication buses or signal lines. Those skilled in the art can understand that the hardware structure shown in Figure 1 does not constitute a limitation on the terminal equipment. End device 300 may include more or fewer components than illustrated, or combinations of certain components, or different component arrangements.

The following is a detailed introduction to each component of the terminal device 300 in conjunction with Figure 3a:

The application processor 301 is the control center of the terminal device 300 and connects various components of the terminal device 300 using various interfaces and buses. In some embodiments, processor 301 may include one or more processing units.

Computer programs are stored in the memory 305, such as the operating system 361 and application programs 363 shown in Figure 3a. The application processor 301 is configured to execute the computer program in the memory 305 to implement functions defined by the computer program. For example, the application processor 301 executes the operating system 361 to implement various functions of the operating system on the terminal device 300 . The memory 305 also stores other data in addition to computer programs, such as data generated during the operation of the operating system 361 and the application program 363. Memory 305 is a non-volatile storage medium, generally including internal memory and external memory. Memory includes but is not limited to random access memory (Random Access Memory, RAM), read-only memory (Read-Only Memory, ROM), or cache (cache), etc. External storage includes but is not limited to flash memory, hard disk, optical disk, universal serial bus (USB) disk, etc. Computer programs are usually stored on external memory, and the processor loads the computer program from external memory into memory before executing it.

The memory 305 can be independent and connected to the application processor 301 through a bus; the memory 305 can also be integrated with the application processor 301 into a chip subsystem.

The MCU 303 is a co-processor used to obtain and process data from the sensor 314. The MCU 303 has less processing power and power consumption than the application processor 301, but has the feature of "always on" and can be used in the application processor. When the 301 is in sleep mode, it continues to collect and process sensor data, ensuring the normal operation of the sensor with extremely low power consumption. In one embodiment, MCU 303 may be a sensor hub chip. Sensors 314 may include light sensors, motion sensors. Specifically, the light sensor can include an ambient light sensor and a proximity sensor. The ambient light sensor can adjust the brightness of the display 351 according to the brightness of the ambient light. The proximity sensor can turn off the power of the display screen when the terminal device 300 moves to the ear. . As a type of motion sensor, the accelerometer sensor can detect the magnitude of acceleration in various directions (usually three axes), and can detect the magnitude and direction of gravity when stationary; the sensor 314 can also include a gyroscope, a barometer, a hygrometer, Other sensors such as thermometers and infrared sensors will not be described in detail here. The MCU 303 and the sensor 314 can be integrated on the same chip, or they can be separate components connected through a bus.

Modem 307 and radio frequency module 309 constitute the terminal equipment 300 communication subsystem, which is used to implement the main functions of wireless communication standard protocols such as 3GPP and ETSI. Among them, Modem 307 is used for encoding and decoding, signal modulation and demodulation, equalization, etc. The radio frequency module 309 is used for receiving and transmitting wireless signals. The radio frequency module 309 includes but is not limited to an antenna, at least one amplifier, coupler, duplexer, etc. The radio frequency module 309 cooperates with the Modem 307 to implement wireless communication functions. Modem 307 can be used as a separate chip or can be combined with other chips or circuits to form a system-on-chip or integrated circuit. These chips or integrated circuits can be applied to all terminal devices that implement wireless communication functions, including: mobile phones, computers, notebooks, tablets, routers, wearable devices, cars, home appliances, etc.

The terminal device 300 can also use the Wi-Fi module 311, the Bluetooth module 313, etc. to perform wireless communication. The Wi-Fi module 311 is used to provide the terminal device 300 with network access that complies with Wi-Fi related standard protocols. The terminal device 300 can access the Wi-Fi access point through the Wi-Fi module 311 and then access the Internet. In some other embodiments, the Wi-Fi module 311 can also serve as a Wi-Fi wireless access point and can provide Wi-Fi network access for other terminal devices. The Bluetooth module 313 is used to implement short-distance communication between the terminal device 300 and other terminal devices (such as mobile phones, smart watches, etc.). The Wi-Fi module 311 in the embodiment of the present application may be an integrated circuit or a Wi-Fi chip, etc., and the Bluetooth module 313 may be an integrated circuit or a Wi-Fi chip. Bluetooth chip, etc.

The positioning module 350 is used to determine the geographical location of the terminal device 300 . It can be understood that the positioning module 350 may be a receiver of a global positioning system (GPS) or a positioning system such as the Beidou Satellite Navigation System or the Russian GLONASS.

The Wi-Fi module 311, the Bluetooth module 313 and the positioning module 350 can be separate chips or integrated circuits, or they can be integrated together. For example, in one embodiment, the Wi-Fi module 311, the Bluetooth module 313 and the positioning module 350 can be integrated on the same chip. In another embodiment, the Wi-Fi module 311, the Bluetooth module 313, the positioning module 350 and the MCU 303 can also be integrated into the same chip.

Input/output devices 335 include but are not limited to: display 351, touch screen 353, audio circuit 355, and so on.

Among them, the touch screen 353 can collect touch events on or near the user of the terminal device 300 (such as the user's operations on the touch screen 353 or near the touch screen 353 using any suitable object such as a finger or a stylus), and Send the collected touch events to other devices (eg, application processor 301). The user's operation near the touch screen 353 can be called floating touch; through floating touch, the user can select, move or drag targets (such as icons, etc.) without directly touching the touch screen 353 . In addition, touch screens 353 can be implemented using various types such as resistive, capacitive, infrared, and surface acoustic wave.

The display (also called a display screen) 351 is used to display information input by the user or information presented to the user. Displays can be configured in the form of LCD screens, organic light-emitting diodes, etc. The touch screen 353 can be overlaid on the display 351. When the touch screen 353 detects a touch event, it is sent to the application processor 301 to determine the type of the touch event. Then the application processor 301 can provide corresponding information on the display 351 according to the type of the touch event. Visual output. Although in Figure 3a, the touch screen 353 and the display 351 are used as two independent components to implement the input and output functions of the terminal device 300, in some embodiments, the touch screen 353 and the display 351 can be integrated to implement the input and output functions of the mobile phone 300. Input and output functions. In addition, the touch screen 353 and the display 351 can be configured in the form of a full panel on the front of the terminal device 300 to achieve a frameless structure.

The audio circuit 3355, speaker 336, and microphone 317 can provide an audio interface between the user and the terminal device 300. The audio circuit 309 can transmit the electrical signal converted from the received audio data to the speaker 313, and the speaker 313 converts it into a sound signal for output; on the other hand, the microphone 314 converts the collected sound signal into an electrical signal, and the audio circuit 309 After receiving, it is converted into audio data, and then the audio data is sent to another terminal device through the Modem 307 and the radio frequency module 309, or the audio data is output to the memory 305 for further processing.

In addition, the terminal device 300 may also have a fingerprint recognition function. For example, the fingerprint collection device may be configured on the back of the terminal device 300 (for example, below the rear camera), or the fingerprint collection device may be configured on the front of the terminal device 300 (for example, below the touch screen 353). For another example, a fingerprint collection device can be configured in the touch screen 353 to implement the fingerprint identification function, that is, the fingerprint collection device can be integrated with the touch screen 353 to implement the fingerprint identification function of the terminal device 300 . In this case, the fingerprint collection device is configured in the touch screen 353, may be a part of the touch screen 353, or may be configured in the touch screen 353 in other ways. The main component of the fingerprint collection device in the embodiment of the present application is a fingerprint sensor. The fingerprint sensor can use any type of sensing technology, including but not limited to optical, capacitive, piezoelectric or ultrasonic sensing technology.

Further, the operating system 361 installed on the terminal device 300 may be or other operating systems, the embodiments of this application do not impose any restrictions on this.

to carry Taking the terminal device 300 of the operating system as an example, as shown in Figure 3a, the terminal device 300 can be logically divided into a hardware layer, an operating system 361, and an application layer. The hardware layer includes the hardware processor 301, microcontroller unit 305, Modem 307, Wi-Fi module 311, sensor 314, positioning module 350 and other hardware resources as mentioned above. The application layer includes one or more applications, such as application 363. The application 363 can be any type of application such as a social application, an e-commerce application, or a browser. As a software middleware between the hardware layer and the application layer, the operating system 361 is a computer program that manages and controls hardware and software resources.

In one embodiment, operating system 361 includes a kernel, a hardware abstraction layer (HAL), libraries and runtime, and a framework. Among them, the kernel is used to provide underlying system components and services, such as power management, memory management, thread management, hardware drivers, etc.; hardware drivers include Wi-Fi drivers, sensor drivers, positioning module drivers, etc. The hardware abstraction layer encapsulates the kernel driver, provides an interface to the framework, and shields low-level implementation details. The hardware abstraction layer 25 runs in user space, while the kernel driver runs in kernel space.

Libraries and runtime are also called runtime libraries, which provide the library files and execution environment required for executable programs to run. In one embodiment, the libraries and runtimes include Android Runtime (ART), libraries, and scenario package runtimes. ART is a virtual machine or virtual machine instance that can convert an application's bytecode into machine code. A library is a program library that provides support for executable programs at runtime, including browser engines (such as webkit), script execution engines (such as JavaScript engines), graphics processing engines, etc. The scene package runtime is the running environment of the scene package, which mainly includes the page execution environment (page context) and the script execution environment (script context). Among them, the page execution environment parses the page code in html, css and other formats by calling the corresponding library, and the script The execution environment parses and executes code or executable files implemented in scripting languages such as JavaScript by calling the corresponding function library.

The framework is used to provide various basic public components and services for each application in the application layer, such as window management, location management, etc. In one embodiment, the framework may include geofencing services, policy services, notification managers, and the like.

The above-described functions of each component of the operating system 361 can be implemented by the application processor 301 executing programs stored in the memory 305 .

It should be noted that the embodiments of the present application can be implemented based on a certain solver, and the solver can be deployed in the aforementioned data processing equipment or user equipment (terminal equipment). In order to facilitate the introduction of this solver, a solver capable of handling linear programming problems in mixed integer programming is schematically introduced below. Figure 3b is an architectural schematic diagram of a solver provided by an embodiment of the present application). As shown in Figure 3b, the solver includes: an equation building module, a preprocessing (presolve) module, a generation (crash) module, a pre-solving module, and a post-processing module. Processing (postsolve) module and recovery (clean up) module.

Specifically, the equation building module can construct a mixed integer programming equation based on the task to be processed. The preprocessing module can structurally simplify the mixed integer programming equation and obtain the preprocessed mixed integer programming equation. The generation module can generate initial points for solution based on the preprocessed mixed integer programming equation. The pre-solving module can use the dual simplex method to solve the pre-processed mixed integer programming equation starting from this initial point, and obtain the optimal solution of the pre-processed mixed integer programming equation. The post-processing module can convert the optimal solution of the preprocessed mixed integer programming equation into an approximate solution of the original mixed integer programming equation. If the restoration module determines that the approximate solution is not the optimal solution of the original mixed integer programming equation, it restores the approximate solution to the optimal solution of the original mixed integer programming equation.

It is worth noting that the computer task processing method or the production plan acquisition method provided by the embodiments of the present application can be implemented through the architecture of the above solver, and the preprocessing module in the solver has been improved to enable the original hybrid integer Planning equations implement better preprocessing to facilitate equation solving. In order to further understand the method provided by the embodiment of the present application, the method will be further introduced below in conjunction with Figure 4. Figure 4 is a schematic flow chart of a computer task processing method provided by an embodiment of the present application. As shown in Figure 4, the method includes:

401. Obtain a mixed integer programming equation. The mixed integer programming equation is used to describe the task to be processed.

In this embodiment, when a certain task needs to be processed, a mixed integer programming equation describing the task to be processed can be obtained. Mixed integer programming equations usually include two major types of equations, namely linear programming equations and nonlinear programming equations. For example, assuming that the task to be processed is to obtain a production plan to be solved for a certain factory (which can also be understood as a production scheduling plan), the production plan to be solved for the factory must not only satisfy the maximization of the demand of the factory and the total number of the factory. Goals such as cost minimization also need to meet the hard constraints and soft constraints specified by the actual business scenario, etc. Then, the mathematical model abstracted based on the factory's production plan to be solved and the requirements that the generated plan needs to meet can be simplified into the corresponding linear programming equation, which is shown in formula (1):
^xx
Ax＝b
x≥0. (1)

In the above formula, x is the variable of the linear programming equation (that is, the production plan to be solved for the factory mentioned above), and is a matrix with m rows and 1 column (each element in this matrix can be regarded as each quantity to be solved in the production plan , for example, this matrix contains 10 elements, which are the amount of raw materials that the factory needs to produce every day within 10 days, etc., and these 10 elements can be either integers or non-integers); A is linear programming The constraint matrix of the equation (including the aforementioned constraints that need to be satisfied by the production plan to be solved for the factory, such as material constraints, productivity constraints, time constraints, traffic constraints, etc.) is a matrix with h rows and m columns, and the constraint matrix A Each row in can be regarded as a constraint that the factory needs to satisfy; c is a matrix corresponding to the number of elements in the variable, also a matrix with m rows and 1 column; b is a matrix corresponding to the number of constraints, with h rows and 1 column matrix. Among them, h and m are both integers greater than or equal to 1.

It should be noted that tasks in different fields can be described by mixed integer programming equations, such as factory production planning in the supply chain field, vehicle route optimization in the logistics field, and facility location in the cloud technology field. , grid dispatching in the electric power field, etc., these tasks can be described by mixed integer programming equations. Different mixed integer programming equations used to describe tasks in different fields often have different variables and different constraint matrices.

402. Determine the target operator in the operator resource pool based on information related to the structure of the mixed integer programming equation. The information is related to at least one of the following: variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation.

After the mixed integer programming equation is obtained, the structure of the mixed integer programming equation can be identified, thereby obtaining information related to the structure of the mixed integer programming equation (it can also be called the structure information of the mixed integer programming equation, which can be used to indicate the structure of the mixed integer programming equation. structure). It is worth noting that the operation of structure identification mainly focuses on analyzing the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation, thereby determining the structural composition of the mixed integer programming equation (that is, the structural composition of the task to be processed) , so the obtained information related to the structure of the mixed integer programming equation is associated with at least one of the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation.

Specifically, since the information related to the structure of the mixed integer programming equation is associated with at least one of the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation, the information related to the structure of the mixed integer programming equation may be Contains at least one of the following: (1) The dimension of the variable of the mixed integer programming equation, that is, the number of elements contained in the variable of the mixed integer programming equation, for example, the number of elements contained in the variable x of the linear programming equation. (2) The sparsity of the constraint matrix of the mixed integer programming equation, that is, the number of elements with a value of zero in the constraint matrix of the mixed integer programming equation, for example, the number of elements with a value of 0 in the constraint matrix A of the linear programming equation quantity. (3) The block information of the constraint matrix of the mixed integer programming equation. The block information of the mixed integer programming equation is used to indicate the submatrix whose elements are zero contained in the constraint matrix of the mixed integer programming equation. For example, assuming a linear programming equation The first three rows of the constraint matrix A are all 0, then the first three rows of the constraint matrix A can be regarded as a sub-matrix, and the block information of the linear programming equation is used to indicate the sub-matrix. (4) The rows in the constraint matrix of the mixed integer programming equation have a calculation relationship with the variables of the mixed integer programming equation. Since each row in the constraint matrix of the mixed integer programming equation will be multiplied by the variables of the mixed integer programming equation, if the mixed integer programming equation is mixed If (all the elements of) a certain row in the constraint matrix of the integer programming equation are zero, it means that there is no calculation relationship between the row and the variables of the mixed integer programming equation. If any of the rows (of any elements) in the constraint matrix of the mixed integer programming equation An element) is not zero, indicating that there is a computational relationship between the row and the variables of the mixed integer programming equation. For example, assuming that the first three rows of the constraint matrix A of the linear programming equation are all 0, then there is no calculation relationship between these three rows and the variable x of the linear programming equation.

After obtaining information related to the structure of the mixed integer programming equation, the target operator can be determined from the operator resource pool based on this information. It should be noted that the operator resource pool contains multiple operators that can be used to implement preprocessing, as shown in Table 1:

Table 1

Based on Table 1, it can be seen that the operator resource pool includes but is not limited to multiple optional operators shown in Table 1. Therefore, the operator resource pool can be selected from the operator resource pool based on the information related to the structure of the mixed integer programming equation. At least one operator is used to preprocess the mixed integer programming equation. This selected operator can be later called the target operator.

Specifically, the target operator can be obtained in the following ways:

(1) In the equation resource pool, determine the preset equation that matches the information related to the structure of the mixed integer programming equation. It should be noted that the equation resource pool contains multiple preset equations with different structures (can also be understood as equation templates with different structures set in advance), and the information related to the structure of the mixed integer programming equation is used to indicate the mixed integer programming The structural category of the equation, therefore, is to traverse the equation resource pool to find a preset equation that matches the information related to the structure of the mixed integer programming equation. Then, the preset equation and the mixed integer programming equation have a similar structure or The same two equations.

(2) In the operator resource pool, determine the operator used for preprocessing of the preset equation as the target operator. It should be noted that after determining the preset equation with a similar or identical structure to the mixed integer programming equation, since the preset equation is set in advance, it can be determined that the preprocessing of the preset equation has used the operator resource pool Which operators are determined and these operators are determined as operators used to implement preprocessing of mixed integer programming equations, that is, target operators. It is worth noting that there are usually multiple target operators. In the preprocessing for the preset equation, each target operator can be recycled multiple times, so it can be determined that each target operator is used for the preset equation. The number of iterations in preprocessing. Then, the multiple target operators can be divided into multiple operator sets according to the number of iterations, and each operator set contains at least one target operator among the multiple target operators. Among these multiple operator sets, the number of iterations of the target operator in the first operator set is 1, the number of iterations of the target operator in the second operator set is 2,..., the kth The number of iterations of the target operator in the operator set is k times, and k is an integer greater than or equal to 1.

It should be understood that this embodiment is only schematically illustrated by associating information related to the structure of the mixed integer programming equation with at least one of the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation. In practical applications, Information related to the structure of the mixed integer programming equation may also be associated with other terms of the mixed integer programming equation, for example, a matrix corresponding to the number of elements in the variables, a matrix corresponding to the number of constraints, etc.

Further, based on the above example, the information related to the structure of the mixed integer programming equation may also include at least one of the following: the sparsity of the matrix corresponding to the number of elements in the variable, the sparsity of the matrix corresponding to the number of constraints, and so on.

It should also be understood that the structure of the preset equation is similar or the same as the structure of the mixed integer programming equation, which usually means that the constraint matrix of the preset equation is similar or the same as the constraint matrix of the mixed integer programming equation.

Furthermore, the structure of the preset equation and the structure of the mixed integer programming equation may also include: the matrices corresponding to the number of elements in the variables of the two equations are similar or the same, and/or the AND constraints of the two equations The matrices corresponding to the quantities are similar or the same, etc.

403. Merge the rows that are in a multiple relationship in the constraint matrix, or set the columns that are in a multiple relationship in the constraint matrix to the same column to obtain a constraint matrix after redundancy is removed.

After the target operator is determined, preprocessing of the mixed integer programming equation can be implemented based on the target operator to obtain the preprocessed mixed integer programming equation. In order to achieve better preprocessing, the mixed integer programming equation can be deredundant in advance, which includes:

(1) In the constraint matrix of the mixed integer programming equation, the rows with multiple relationships can be determined in a certain way, and these rows can be merged to obtain the constraint matrix after the merged rows. Still as in the above example, assuming a ₁ and a ₂ are the first and second rows of the constraint matrix A respectively, then the following calculations can be performed for a ₁ and a ₂ respectively:

In the above formula, a ₁₁ is the first element of a ₁ , a ₁₂ is the second element of a ₁ , ..., a _1m is the m-th element of a ₁ , similarly, a ₂₁ , a ₂₂ and a _2m , which will not be described again here. λ and λ` are both preset parameter matrices, and each element in these two matrices is a preset value.

Based on formula (2), a ratio can be set for each row of the constraint matrix A, where the ratio of the x-th row (i.e. ratioox) = sum`[x]/sum[x], x=1,...,h. If ratiox = ratioy, it means that a <sub>x</sub> and a <sub>y</sub> may be in a multiple relationship, that is, a <sub>x</sub> = qa <sub>y</sub> , which means that a <sub>x</sub> and a <sub>y</sub> are potential parallel lines. Then, the ratios of all rows in the constraint matrix A can be calculated, and the ratios of all rows are sorted by size (at the same time, all rows in the constraint matrix A are also sorted by the size of the ratio). The same ratio corresponds to Rows are potential parallel rows. Then, further judgment can be made on the potential parallel rows to determine whether these potential parallel rows are real parallel rows (that is, whether these potential parallel rows are really multiples). After determining the real parallel rows, you can delete For rows with larger elements, retain the rows with the smallest elements (of course, you can also delete rows with smaller elements, retain the rows with the largest elements, etc.), thereby merging the parallel rows of the constraint matrix A to obtain the constraint matrix A after the merged rows. .As shown in the example of formula (4), after sorting the ratios of all rows of matrix A, ratio1=sum`[1]/sum[1], ratio2=sum`[2]/sum[2 ],..., ratioh=sum`[h]/sum[h]. Assume ratio1=ratio2, indicating that a ₁ and a ₂ may be in a multiple relationship, that is, a ₁ may be equal to qa ₂ , that is to say, a ₁ and a ₂ are Potential parallel rows. Then, we can further determine whether a ₁ and a ₂ are really in a multiple relationship. If they are in a multiple relationship, it means that a ₁ and a ₂ are truly parallel rows, and the smaller row can be retained. If the two are in a multiple relationship, If the two are not in a multiple relationship, no operation will be performed on them.

(2) In the constraint matrix after merging rows, the columns with multiple relationships can be determined in the same way, and these columns are set to the same columns to obtain the constraint matrix after redundancy removal (which can also be called the transformed constraint matrix). Still as in the above example, let A ₁ and A ₂ be respectively the first and second columns of the constraint matrix A after merging rows. Similarly, a ratio can be set for each column of the constraint matrix A after merging rows, based on ratio of all rows to determine the real parallel columns (for the process of obtaining real parallel columns, please refer to the aforementioned process of obtaining real parallel rows, which will not be repeated here). After determining the real parallel columns, these parallel columns can be Set to one of the columns to obtain the constraint matrix A after redundancy removal. For example, assuming that A ₁ and A ₂ are truly parallel columns, A ₁ = qA ₂ , and transforming the elements of the first row in the variable x from x ₁ to qx ₁ can make the constraint matrix after merging the rows The first column in A is transformed from A ₁ to A ₂ , as shown in formula (3):
min c ₁ x ₁ +c ₂ x ₂ +…＝c ₁ /q*(qx ₁ )+c ₁ x ₂ +…
stx ₁ A ₁ +x ₂ A ₂ +…＝(qx ₁ +x ₂ )A ₂ +… (3)

Based on the above formula, it can be seen that when both the constraint matrix A and the variable x are transformed as in formula (3), the element in the first row of the matrix c corresponding to the number of elements in the variable is transformed from c ₁ to c ₁ /q. Correspondingly, the value range of x ₁ needs to be changed:

In the above formula, clo ₁ is the maximum value of x ₁ , and cup ₁ is the minimum value of x ₁ .

404. Update the mixed integer programming equation based on the redundantly removed constraint matrix to obtain an updated mixed integer programming equation.

After obtaining the constraint matrix after removing redundancy, the original mixed integer programming equation can be updated with the constraint matrix after removing redundancy to obtain the updated mixed integer programming equation. Based on the aforementioned formula (4), it can be seen that in the original mixed integer programming equation, after replacing the original constraint matrix with the constraint matrix after removing redundancy, the original variables also need to be replaced with the transformed variables accordingly. The original The matrix corresponding to the number of elements in the variable also needs to be replaced with the transformed matrix corresponding to the number of elements in the variable, so as to obtain the updated mixed integer programming equation.

It should be understood that steps 403 and 404 are optional steps. In practical applications, after the target operator is determined, the mixed integer programming equation does not need to be deredundant and updated (that is, steps 403 and 404 are not performed). Then, the processing object in step 405 is the original mixed Integer programming equations.

405. Preprocess the updated mixed integer programming equation based on the target operator to obtain the preprocessed mixed integer programming equation.

After the updated mixed integer programming equation is obtained, the objective operator can be used to implement preprocessing for the updated mixed integer programming equation, and the preprocessed mixed integer programming equation is obtained.

Specifically, the updated mixed integer programming equation can be preprocessed through the following iteration strategy:

Since k operator sets have been obtained, in this embodiment, the operations performed on each operator set are similar. Therefore, the kth operator set is used as an example for a schematic introduction below. Assume that the kth operator set contains n target operators (n is an integer greater than or equal to 1). Then, the number of iterations of these n target operators is k times. Then, the updated mixed integer programming equation can be Execute the steps shown in Figure 5 (Figure 5 is a schematic flowchart of preprocessing provided by the embodiment of this application):

(1) In the j-th iteration of n target operators, before executing the i-th operator on the updated mixed integer programming equation, the index of the i-th target operator in the first iteration can be detected first. Whether the sum of indicators of the i-th target operator in the j-1th iteration is less than the first threshold (the first threshold is usually determined based on the number of iterations k, generally with the number of iterations k increases, that is, there is a positive correlation between the two). Among them, i=1,...,n, j=1,...,k.

(2) If the sum of the index of the i-th target operator in the 1st iteration to the index of the i-th target operator in the j-1th iteration is less than the first threshold, it means that the i-th target operator needs to continue to be executed thereafter. target operator, then further detect whether the index of the i-th target operator in the j-1th iteration is less than the second threshold (the second threshold is usually determined based on the number of iterations k, and generally increases as the number of iterations k increases. increases, that is, there is a positive correlation between the two).

(3) If the sum of the index of the i-th target operator in the 1st iteration to the index of the i-th target operator in the j-1th iteration is greater than or equal to the first threshold, it means that there is no need to continue thereafter. If the i-th target operator is executed, the i-th target operator will be skipped in the j-th iteration to the k-th iteration.

(4) If the index of the i-th target operator in the j-1th iteration is less than the second threshold, it means that the i-th operator can currently be executed, and the i-th target operator is executed on the updated mixed integer programming equation. operator, and generate the index of the i-th target operator in the j-th iteration. It is worth noting that the index of the i-th target operator in the j-th iteration can be generated in the following way: first, determine the rows deleted by the i-th target operator in the 1st iteration to the j-th iteration. in the updated mixed integer programming equation. Then, calculation is performed based on this proportion to obtain the index of the i-th target operator in the j-th iteration.

(5) If the index of the i-th target operator in the j-1th iteration is greater than or equal to the second threshold, it means that the i-th operator cannot be executed currently, and then jump to the i+1-th target operator. And perform the above steps for the i+1th target operator until the jth iteration of the n target operators is completed.

(6) Continue to execute the j+1th iteration of n target operators (that is, repeat steps (1) to step (5)) until the kth iteration of n target operators is completed, and obtain the preprocessed Mixed integer programming equations.

It should be understood that in the kth operator set, the n target operators have been sorted in a certain order (the order is the execution order of the n target operators in the preprocessing of the preset equation), so in the In the updated mixed integer programming equation, these n operators can be executed in this order in each iteration.

It should be understood that for other operator sets except the kth operator set, the same operations as those performed for the kth operator set can also be performed, which will not be described in detail here, thereby obtaining the preprocessed mixed integer programming. equation.

406. Solve the preprocessed mixed integer programming equation, and the obtained solution is used as the processing result of the task.

After obtaining the preprocessed mixed integer programming equation, the preprocessed mixed integer programming equation can be solved to obtain the corresponding solution, which can be used as the processing result of the aforementioned task.

In order to further understand the above process, the computer task processing method provided by the embodiment of the present application is further introduced below in conjunction with multiple application examples. In the first application example, assuming that the task to be processed is the production scheduling plan of a certain factory in the supply chain field, Figure 6 shows the role of this production planning in the entire supply chain (Figure 6 shows the implementation of this application An application example diagram of the computer task processing method provided in the example), as shown in the production plan module in Figure 6. From the perspective of mathematical models, this problem is a typical large-scale, multi-objective, sparsely constrained, and multi-variable allocation problem. The input of this problem is the customer's order demand and existing processing resources (raw materials, production capacity, etc.). The output is the processing production plan (processing volume, transportation volume, delivery volume, etc.) arranged according to the order demand. The expected output result is the ability to Minimize processing cost, processing time and maximize order fulfillment rate (that is, deliver customer needs on time and quantity), etc., and all processing constraints need to be met.

In the problem of obtaining the production schedule, the production schedule needs to satisfy different constraints, including material constraints, transportation constraints, inventory constraints, production capacity upper limit constraints, etc. For example: customer demand is to manufacture 2,000 PCs and 1,000 hosts within three days. 800 laptops. Figure 7 is a schematic diagram of the bill of material (BOM) in the production process provided by the embodiment of the present application. As shown in Figure 7, the PC generation process is taken as an example to illustrate how the BOM affects the production process: It is assumed that each PC needs to be configured with a host, and each laptop can be configured individually. There are two factories A and B. Factory A can process PCs and mainframes, while factory B can process mainframes and laptops. The daily processing capacity of both is limited, that is, factory A can only process up to 1,000 PCs or mainframes per day. Factory B can only process up to 1,000 hosts or laptops per day. Then, a legal and better production plan is: on the first day, factory A arranges to produce 1,000 hosts, and factory B arranges to produce 1,000 hosts. After factory B completes production, the 1,000 hosts are sent to factory A; on the second day, factory A arranges to produce 1,000 hosts. On the third day, Factory A uses the 1,000 hosts it has already produced to complete the assembly and production of 1,000 PCs, while Factory B continues to arrange the production of 1,000 hosts; on the third day, Factory A uses the 1,000 hosts sent by Factory B on the first day to complete the assembly and production. For the assembly of 1,000 PCs, Factory B arranges to produce 800 laptops. According to the above arrangement, the two factories will complete all customer order requirements on the third day and comply with various constraints such as materials, productivity, time, and transportation. The above example does not take into account multiple conflicting objectives (cost, order delivery fulfillment rate, etc.), and real industrial scenarios are far more complex than the example problem.

Then, the factory’s production plan to be solved can be obtained through the following steps:

Step 1: Construct the equation. According to business requirements, set up a mixed integer programming equation. This equation can include mathematical relationships between variables and constraint matrices. The variables are the factory's production plan to be solved, and the constraint matrix is the processing required for the production plan to be solved. Constraints, processing constraints can include at least one of the following: material constraints, productivity constraints, time constraints or traffic constraints, etc., where material constraints can include the upper limit of materials that can be used in production, material ratios, etc. Productivity constraints can include The upper limit of the number of production workers required during production, the upper limit of work efficiency, etc. The time constraints can include the delivery date of the product, etc. The transportation constraints can include the upper limit of the transportation cost required during production and the upper limit of the number of transportation vehicles required during production. wait. It can be seen that solving this equation is equivalent to obtaining the production plan of the factory.

Step 2: Read the constructed mixed integer programming equation into the solver, call the specified interface of the solver, and perform preprocessing of the mixed integer programming equation. This preprocessing process can be divided into the following steps:

(1) Problem structure identification: Perform structure identification on the mixed integer programming equation to obtain information related to the structure of the mixed integer programming equation.

(2) Delete redundancy: For the constraint matrix of the mixed integer programming equation, determine the parallel rows and rows, so as to achieve corresponding redundancy processing for parallel rows and rows. In the process of removing redundancy, the mixed integer programming equation can be traversed according to the upper and lower dependencies of the BOM, which will bring better simplification effects. The main reason is that the solution of the upper-level variables will strictly depend on the lower-level variables. In terms of space, it will be more convenient to simplify the upper-level variables by processing the lower-level variables first.

(3) Operator selection: Match the target operator according to the structural information obtained from the structure recognition, open the execution parameters of each target operator (for example, the number of iterations, iteration order, adaptive indicators, etc.), and select the target operator according to a certain strategy. Mixed integer programming equations implement these objective operators to achieve preprocessing of the equations. The goal to be achieved by this strategy is: if the indicator shows that the operator cannot simplify the variables and constraint matrices after a certain number of executions, the execution will stop early.

(4) Continuous simplification: The above process can be re-executed, that is, steps (1) to (3), that is, preprocessing is performed multiple times until the equation meets the preset convergence judgment conditions, and the preprocessed mixed integer programming is obtained. equation.

It should be noted that sub-step (2) in step 2 is an optional step. In actual applications, it may or may not be executed. The selection can be made according to actual needs and is not limited here.

Step 3: Method selection. Choose from different methods in the algorithm library, including primitive simplex method, dual simplex method, interior point method, etc.

Step 4: Based on the method selected in step 3, complete the solution of the preprocessed mixed integer programming equation.

In addition, in the multi-factory scheduling problem in the supply chain field, the solver provided by the embodiment of the present application and the solver of the related technology can also be compared. The comparison result is shown in Figure 8 (Figure 8 is provided by the embodiment of the present application A schematic diagram of the comparison results). By comparing the calculation results using the related technology and the calculation results using the embodiments of the present application, it can be seen that compared with the related technology 1, the present invention can reduce the solution time by more than 18% end-to-end. Regarding related technology 2 and related technology 3, the present invention can reduce the solving time by more than 60%.

In the second application example, assuming that the task to be processed is the solution of a certain public data set, the solver provided by the relevant technology and the solver provided by the embodiment of the present application can be used to respectively solve the mixed integer programming equation describing the task. Solve and compare. The comparison results are shown in Table 2:

Table 2

Based on Table 2, it can be seen that compared with related technology 1, the solving time of the embodiment of the present application can be reduced from 1394s to less than 100s, surpassing the solving time of related technology 2 and related technology 3. It can be seen that in the solver of the embodiment of the present application Preprocessing module with excellent performance.

In the third application example, the task to be processed is the vehicle route optimization problem in the logistics field. The vehicle routing problem was first proposed in 1959. The problem is to designate a certain number of customers, each with different quantities of goods requirements. The distribution center provides goods to the customers, and a fleet is responsible for distributing the goods and organizing appropriate driving routes. The goal is to satisfy customer needs and achieve goals such as the shortest distance, the lowest cost, and the least time consuming under certain constraints.

This problem requires finding the optimal solution among a limited set of feasible solutions. This type of problem requires finding the optimal value (maximum or minimum value) of the objective function that satisfies the given constraints on a discrete and finite mathematical structure. Problems, also called discrete optimization problems, are all related to order.

Different from the multi-factory production scheduling problem, this problem requires more consideration of the loading capacity of different types of vehicles. Therefore, when building a mathematical model (that is, a mixed integer programming equation used to describe the problem), it is necessary to consider the loading capacity of different types of vehicles. There is a time sequence consideration. It should be noted that in the mixed integer programming equation describing this problem, the variable is the vehicle adoption plan to be solved, and different elements in this variable represent different types of vehicles (since vehicle paths are often difficult to express numerically, so here a certain refers to the number of vehicles traveling along a certain route), and the constraint matrix contains one or more of the capacity constraints, product constraints, time constraints and road condition constraints that the vehicle adoption plan to be solved needs to satisfy, where , the capacity constraint includes the upper capacity limit of the vehicle during transportation, etc., the product constraint includes the product categories that the vehicle can transport during transportation, etc., the time constraint includes the time sequence of different types of vehicles arriving at the destination, etc., and the road condition constraint includes different types of vehicles. The road conditions that the vehicle can travel on, etc.

Therefore, in this type of problem, there is often an obvious mathematical structure in the time dimension of the variable. Therefore, the preprocessing method provided by the embodiment of the present application can be relied on to simplify the equation, thereby accelerating the solution of the equation.

In the fourth application example, the task to be processed is the center location problem in the field of cloud technology. In practical applications, enterprises Companies need to consider rival competition when making location decisions. Generally speaking, users need to consider a discrete network where two service providers (leader and follower) open a certain number of facilities one after another to compete for market share. Each user seeks services from the nearest facility.

Different from the previous example, this application case requires solving a two-level linear programming equation, where the upper-level problem represents the leader's decision-making, and its goal is maximum coverage, while the lower-level problem is the corresponding NP-hard problem for the follower. Assuming that the follower adopts a greedy strategy, the follower's response can be integrated into the constraints of the leader's problem.

Considering the difference between the two layers, when using the preprocessing method provided by the embodiment of this application, the execution process needs to be executed from the bottom layer to the upper layer, and the aforementioned iteration strategy (implementing early stopping of certain operators, etc.) needs to be implemented, so as to more Further reduce the time-consuming preprocessing.

It should be noted that the variable of the mixed integer programming equation constructed based on the central location problem is the address selection plan to be solved. Different elements in the variable represent the addresses of different objects (for example, enterprises, factories, etc.) (for example, , longitude, latitude, etc.), the constraint matrix contains one or more of the distribution constraints, traffic environment constraints and rule constraints that the address selection plan to be solved needs to satisfy, where the distribution constraints include the current distribution of different objects, Future distribution predictions of different objects, business volume growth rates of objects, regional distribution of interactions between objects, etc. Traffic environment constraints include transportation conditions, distribution conditions, land use conditions, etc. when objects interact with each other. Rule constraints include the relationship between objects. regulations that need to be followed for interactions, etc.

In the fifth application example, it is assumed that the task to be processed is a power grid dispatching problem in the electric power field. The power grid dispatching problem mainly includes two stages:

In the first stage, the economic dispatch problem of the non-convex power system is converted into a mixed integer programming equation through linearization processing and mixed integer coding technology, and then an optimization software package is used to solve the mixed integer programming equation to obtain a satisfactory feasible solution;

In the second stage, based on the feasible solution obtained in the first stage, the output range of the unit is compressed, and the mixed integer programming equation is resolved to obtain the final economic dispatch plan of the power system. Therefore, power grid dispatching is a multi-stage mathematical model.

In this application example, the multi-stage model needs to consider more boundary constraints of different time periods. Therefore, in the process of using the preprocessing method provided by the embodiment of this application, it is necessary to consider the segmentation of time periods.

It should be noted that the variable of the mixed integer programming equation constructed based on the power grid dispatch problem is the power consumption dispatch plan to be solved. Different elements in this variable represent the power consumption in different regions, and the constraint matrix contains the power consumption to be solved. One or more of the quantity constraints and price constraints that the electricity consumption dispatch plan needs to satisfy. Among them, the quantity constraints include the upper limit of electricity consumption in each region, etc., and the price constraints include the upper limit of the total price of electricity in each region. etc.

In the embodiment of the present application, after obtaining the mixed integer programming equation used to describe the task to be processed, the target operator can be determined in the operator resource pool based on information related to the structure of the mixed integer programming equation. This information is related to the following At least one item is relevant: the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation. Then, the mixed integer programming equation can be preprocessed based on the target operator to obtain the preprocessed mixed integer programming equation. Finally, the preprocessed mixed integer programming equation can be solved, and the obtained solution can be used as the processing result of the task. In the foregoing process, the variables and constraint matrices of the mixed integer programming equation can be identified, thereby providing information related to the structure of the mixed integer programming equation. Based on this information, the structure of the mixed integer programming equation can be determined, that is, the mixed integer programming equation belongs to Which type of mixed integer programming equation (used to describe a specific class of tasks), so as to select a specific target operator for this mixed integer programming equation, so as to achieve the mixed integer programming equation. The preprocessing of integrated integer programming equations can effectively reduce the time consumed in the equation preprocessing process and improve the performance of the equation solving process.

Furthermore, the embodiment of the present application provides an equation resource pool. After determining which type of preset mixed integer programming equation the mixed integer programming equation belongs to, that is, a certain preset equation in the equation resource pool (which can also be called an equation An equation template in the resource pool), the operator used for preprocessing of the preset equation, the number of operator iterations and the order of operator execution can be used as the operator used for preprocessing of the mixed integer programming equation , the number of operator iterations and the order of operator execution to reduce the impact of human intervention on preprocessing, thereby optimizing the preprocessing effect of the equation.

Furthermore, the embodiments of this application also provide an iterative strategy, that is, during the iterative process, when the role of an operator becomes smaller, it can be considered that the benefit of continuing to execute the operator in subsequent iterations is limited, and it will be stopped in time. Execute this operator. Moreover, a corresponding stopping mechanism is also designed for the preprocessing cycle. In this way, not only the mixed integer programming equation can be effectively simplified, but also unnecessary preprocessing processes can be reduced, thereby reducing the time overhead of preprocessing.

The above is a detailed description of the computer task processing method provided by the embodiment of the present application. The computer task processing device provided by the embodiment of the present application will be introduced below. Figure 9 is a schematic structural diagram of a computer task processing device provided by an embodiment of the present application. The device includes:

The acquisition module 901 is used to obtain the mixed integer programming equation, which is used to describe the task to be processed;

The determination module 902 is used to determine the target operator in the operator resource pool based on the information related to the structure of the mixed integer programming equation, where the target operator is a part of the operators in the operator resource pool;

The preprocessing module 903 is used to preprocess the mixed integer programming equation based on the target operator to obtain the preprocessed mixed integer programming equation;

The solving module 904 is used to solve the preprocessed mixed integer programming equation, and the obtained solution is used as the processing result of the task.

In the embodiment of the present application, after obtaining the mixed integer programming equation used to describe the task to be processed, the target operator can be determined in the operator resource pool based on information related to the structure of the mixed integer programming equation. This information is related to the following At least one item is relevant: the variables of the mixed integer programming equation and the constraint matrix of the mixed integer programming equation. Then, the mixed integer programming equation can be preprocessed based on the target operator to obtain the preprocessed mixed integer programming equation. Finally, the preprocessed mixed integer programming equation can be solved, and the obtained solution can be used as the processing result of the task. In the foregoing process, the variables and constraint matrices of the mixed integer programming equation can be identified, thereby providing information related to the structure of the mixed integer programming equation. Based on this information, the structure of the mixed integer programming equation can be determined, that is, the mixed integer programming equation belongs to Which type of mixed integer programming equation (used to describe a specific type of task), so as to select a specific target operator for the mixed integer programming equation, thereby realizing the preprocessing of the mixed integer programming equation, which can effectively reduce the cost of the equation preprocessing process time consumed and improves the performance of the equation solving process.

In a possible implementation, the information related to the structure of the mixed integer programming equation includes at least one of the following: the dimensions of the variables, the sparsity of the constraint matrix, the block information of the constraint matrix, and the existence of a calculation relationship with the variables in the constraint matrix. rows, where the blocking information is used to indicate that the constraint matrix contains submatrices with zero elements.

In a possible implementation, the determination module 902 is configured to: in the equation resource pool, determine the preset equation that matches the information related to the structure of the mixed integer programming equation; in the operator resource pool, determine for The operator used in the preprocessing of the preset equation is determined as the target operator.

In a possible implementation, in the preprocessing of mixed integer programming equations, the number of target operators is n, And the number of iterations of n target operators is k times, and n and k are both integers greater than or equal to 1. The preprocessing module 903 is used to: in the j-th iteration of n target operators, if the i-th If the sum of the index of the target operator in the 1st iteration to the index of the i-th target operator in the j-1th iteration is greater than or equal to the first threshold, then in the j-th iteration to the k-th iteration, Skip the i-th target operator, i=1,...,n,j=1,...,k; or, if the index of the i-th target operator in the first iteration to the i-th target operator is If the sum of indicators in the j-1th iteration is less than the first threshold, then detect whether the indicator of the i-th target operator in the j-1th iteration is less than the second threshold; or, if the i-th target operator is in If the index in the j-1th iteration is less than the second threshold, execute the i-th target operator on the mixed integer programming equation and generate the index of the i-th target operator in the j-th iteration; or, if the i-th target operator is If the index of the target operator in the j-1th iteration is greater than or equal to the second threshold, then perform the above steps for the i+1th target operator until the jth iteration of n target operators is completed; execute n The j+1th iteration of n target operators is completed until the kth iteration of n target operators is completed, and the preprocessed mixed integer programming equation is obtained.

In a possible implementation, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the number of iterations of the n target operators is the same.

In a possible implementation, the index of the i-th target operator in the j-th iteration is determined based on the proportion of rows deleted by the i-th target operator in the mixed integer programming equation.

In a possible implementation, the first threshold and the second threshold are determined based on the number of iterations of n target operators.

In a possible implementation, the device further includes: a redundancy removal module, used to merge the rows with a multiple relationship in the constraint matrix, or set the columns with a multiple relationship in the constraint matrix to the same column, to obtain The constraint matrix after the redundancy is removed; the update module is used to update the mixed integer programming equation based on the redundant constraint matrix to obtain the updated mixed integer programming equation; the preprocessing module 903 is used to update the mixed integer programming equation based on the target operator pair The updated mixed integer programming equation is preprocessed to obtain the preprocessed mixed integer programming equation.

Figure 10 is a schematic structural diagram of a production plan acquisition device provided by the implementation method of this application. As shown in Figure 10, the device includes:

The acquisition module 1001 is used to obtain the mixed integer programming equation, which is used to describe the production plan to be solved in the supply chain field;

The determination module 1002 is used to determine the target operator in the operator resource pool based on the information related to the structure of the mixed integer programming equation, where the target operator is a part of the operators in the operator resource pool;

The preprocessing module 1003 is used to preprocess the mixed integer programming equation based on the target operator to obtain the preprocessed mixed integer programming equation;

The solving module 1004 is used to solve the preprocessed mixed integer programming equation to obtain the solved production plan.

In the embodiment of the present application, after obtaining the mixed integer programming equation used to describe the production plan to be solved, the target operator can be determined in the operator resource pool based on the information related to the structure of the mixed integer programming equation. This information is related to At least one of the following is relevant: the variables of a mixed integer programming equation and the constraint matrix of a mixed integer programming equation. Then, the mixed integer programming equation can be preprocessed based on the target operator to obtain the preprocessed mixed integer programming equation. Finally, the preprocessed mixed integer programming equation can be solved to obtain the solved production plan. In the foregoing process, the variables and constraint matrices of the mixed integer programming equation can be identified, thereby providing information related to the structure of the mixed integer programming equation. Based on this information, the structure of the mixed integer programming equation can be determined, that is, the mixed integer programming equation belongs to Which type of mixed integer programming equation (used to solve the production plan of factories and enterprises in the field of supply chain), so as to select a specific target operator for this mixed integer programming equation, so as to achieve the preprocessing of this mixed integer programming equation, can be effective Reduce the time consumed by equation preprocessing and Improve the performance of the equation solving process.

In a possible implementation, the variable of the mixed integer programming equation is the production plan to be solved, and the constraint matrix of the mixed integer programming equation includes at least one of the following: material constraints, productivity constraints, and time that the production plan to be solved needs to satisfy. restraints or traffic constraints.

In a possible implementation, the information related to the structure of the mixed integer programming equation includes at least one of the following: the dimensions of the variables, the sparsity of the constraint matrix, the block information of the constraint matrix, or the existence of a computational relationship with the variables in the constraint matrix. rows, where the blocking information is used to indicate that the constraint matrix contains submatrices with zero elements.

In a possible implementation, the determination module 1002 is configured to: in the equation resource pool, determine the preset equation that matches the information related to the structure of the mixed integer programming equation; in the operator resource pool, determine the preset equation for The operator corresponding to the preprocessing of the preset equation is determined as the target operator.

In a possible implementation, in the preprocessing of mixed integer programming equations, the number of target operators is n, and the number of iterations of the n target operators is k times, and both n and k are greater than or equal to 1. Integer of If the sum of indicators in one iteration is greater than or equal to the first threshold, then in the j-th iteration to the k-th iteration, the i-th target operator is skipped, i=1,...,n,j=1,... , k; or if the sum is less than the first threshold, then detect whether the index of the i-th target operator in the j-1th iteration is less than the second threshold; or, if the i-th target operator is in the j-1th iteration If the index in the iteration is less than the second threshold, execute the i-th target operator on the mixed integer programming equation and generate the index of the i-th target operator in the j-th iteration; or, if the i-th target operator is in If the index in the j-1th iteration is greater than or equal to the second threshold, then the above steps are performed for the i+1th target operator until the jth iteration of n target operators is completed; The j+1th iteration is performed until the kth iteration of the n target operators is completed, and the preprocessed mixed integer programming equation is obtained.

In a possible implementation, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the number of iterations of the n target operators is the same.

In a possible implementation, the index of the i-th target operator in the j-th iteration is determined based on the proportion of rows deleted by the i-th target operator in the mixed integer programming equation.

In a possible implementation, the first threshold and the second threshold are determined based on the number of iterations of n target operators. In a possible implementation, the device further includes: a redundancy removal module, used to merge the rows with a multiple relationship in the constraint matrix, or set the columns with a multiple relationship in the constraint matrix to the same column, to obtain The constraint matrix after the redundancy has been removed; the update module is used to update the mixed integer programming equation based on the redundancy removal constraint matrix to obtain the updated mixed integer programming equation; the preprocessing module 1003 is used to update the mixed integer programming equation based on the target operator pair The updated mixed integer programming equation is preprocessed to obtain the preprocessed mixed integer programming equation.

It should be noted that the information interaction, implementation process, etc. between the modules/units of the above-mentioned device are based on the same concept as the method embodiments of the present application, and the technical effects they bring are the same as those of the method embodiments of the present application. The specific content can be Refer to the description in the method embodiments shown above in the embodiments of the present application, which will not be described again here.

Embodiments of the present application also relate to a computer storage medium, which includes computer-readable instructions. When the computer-readable instructions are executed, the method steps of the embodiment shown in Figure 4 are implemented.

Embodiments of the present application also relate to a computer program product containing instructions that, when run on a computer, cause the computer to execute the method steps of the embodiment shown in FIG. 4 .

Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be described again here.

In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or can be integrated into another system, or some features can be ignored, or not implemented. On the other hand, the coupling or direct coupling or communication connection between each other shown or discussed may be through some interfaces, and the indirect coupling or communication connection of the devices or units may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or they may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or software functional units.

If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or contributes to the existing technology, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program code. .

Claims (18)

A method for obtaining a production plan, characterized in that the method includes: Obtain a mixed integer programming equation, which is used to describe the production plan to be solved; Based on information related to the structure of the mixed integer programming equation, determine a target operator in the operator resource pool, where the target operator is a part of the operators in the operator resource pool; Preprocess (presolve) the mixed integer programming equation based on the target operator to obtain a preprocessed mixed integer programming equation; The preprocessed mixed integer programming equation is solved to obtain a solved production plan. The method according to claim 1, characterized in that the variable of the mixed integer programming equation is the production plan to be solved, and the constraint matrix of the mixed integer programming equation includes at least one of the following: the to-be-solved Material constraints, productivity constraints, time constraints or transportation constraints that the production plan needs to meet. The method according to claim 2, characterized in that the information includes at least one of the following: the dimension of the variable, the sparsity of the constraint matrix, the block information of the constraint matrix, or the A row that has a calculation relationship with the variable, wherein the blocking information is used to indicate a submatrix whose elements are zero contained in the constraint matrix. The method according to any one of claims 1 to 3, characterized in that, based on the information related to the structure of the mixed integer programming equation, determining the target operator in the operator resource pool includes: In the equation resource pool, determine a preset equation that matches the information related to the structure of the mixed integer programming equation; In the operator resource pool, the operator corresponding to the preprocessing of the preset equation is determined as the target operator. The method according to claim 4, characterized in that, in the preprocessing of the mixed integer programming equation, the number of the target operators is n, and the number of iterations of the n target operators is k times, n and k are both integers greater than or equal to 1. The mixed integer programming equation is preprocessed based on the target operator, and the preprocessed mixed integer programming equation obtained includes: In the j-th iteration of n target operators, if the index of the i-th target operator in the first iteration is between the index of the i-th target operator in the j-1 iteration The sum is greater than or equal to the first threshold, then in the j-th iteration to the k-th iteration, the i-th target operator is skipped, i=1,...,n, j=1,...,k; or If the sum is less than the first threshold, detect whether the index of the i-th target operator in the j-1th iteration is less than the second threshold; or, If the index of the i-th target operator in the j-1th iteration is less than the second threshold, execute the i-th target operator on the mixed integer programming equation and generate the i-th target operator. indicator in the jth iteration; or, If the index of the i-th target operator in the j-1th iteration is greater than or equal to the second threshold, then perform the above steps for the i+1-th target operator until the n target operators are completed. jth iteration; The j+1th iteration of the n target operators is executed until the kth iteration of the n target operators is completed, and the preprocessed mixed integer programming equation is obtained. The method according to claim 5, characterized in that, in the preprocessing for the preset equation and the preprocessing for the mixed integer programming equation, the number of iterations of the n target operators is the same. The method according to claim 5 or 6, characterized in that the index of the i-th target operator in the j-th iteration is based on the row deleted by the i-th target operator in the mixed integer The proportion in the planning equation is determined. The method according to any one of claims 5 to 7, characterized in that the first threshold and the second threshold The value is determined based on the number of iterations of the n target operators. The method according to any one of claims 1 to 8, characterized in that, before preprocessing the mixed integer programming equation based on the target operator to obtain the preprocessed mixed integer programming equation, Methods also include: Merge the rows in the constraint matrix that are in a multiple relationship, or set the columns in the constraint matrix that are in a multiple relationship to the same column to obtain a constraint matrix after removing redundancy; Update the mixed integer programming equation based on the redundantly removed constraint matrix to obtain an updated mixed integer programming equation; Preprocessing the mixed integer programming equation based on the target operator, and obtaining the preprocessed mixed integer programming equation includes: The updated mixed integer programming equation is preprocessed based on the target operator to obtain a preprocessed mixed integer programming equation. A computer task processing method, characterized in that the method includes: Obtain a mixed integer programming equation, which is used to describe the task to be processed; Based on information related to the structure of the mixed integer programming equation, determine a target operator in the operator resource pool, where the target operator is a part of the operators in the operator resource pool; Preprocess the mixed integer programming equation based on the target operator to obtain a preprocessed mixed integer programming equation; The preprocessed mixed integer programming equation is solved, and the obtained solution is used as the processing result of the task. The method of claim 10, wherein the information includes at least one of the following: dimensions of variables of the mixed integer programming equation, sparsity of the constraint matrix of the mixed integer programming equation, the constraint matrix The blocking information or the rows in the constraint matrix that have a computational relationship with the variable, wherein the blocking information is used to indicate a submatrix whose elements are zero contained in the constraint matrix. The method according to claim 10 or 11, characterized in that, based on the information related to the structure of the mixed integer programming equation, determining the target operator in the operator resource pool includes: In the equation resource pool, determine a preset equation that matches the information related to the structure of the mixed integer programming equation; In the operator resource pool, the operator corresponding to the preprocessing of the preset equation is determined as the target operator. A computer task processing device, characterized in that the device includes: An acquisition module, used to acquire a mixed integer programming equation, where the mixed integer programming equation is used to describe the task to be processed; A determination module, configured to determine a target operator in the operator resource pool based on information related to the structure of the mixed integer programming equation, where the target operator is a part of the operators in the operator resource pool; A preprocessing module, configured to preprocess the mixed integer programming equation based on the target operator to obtain a preprocessed mixed integer programming equation; A solving module is used to solve the preprocessed mixed integer programming equation, and the obtained solution is used as the processing result of the task. A computer task processing device, characterized in that it includes: a processor, a memory, a bus, and an input and output device. The processor is connected to the memory and the input and output devices. The bus is connected to the processor, the memory, and the input and output devices respectively. Input and output devices are connected; The input and output device is used to obtain a mixed integer programming equation, and the mixed integer programming equation is used to describe the processing to be processed tasks; An operator resource pool is stored in the memory; The processor is used for: Obtain an operator resource pool from the memory, and determine a target operator in the operator resource pool based on information related to the structure of the mixed integer programming equation, where the target operator is the operator Some operators in the resource pool; Preprocess the mixed integer programming equation based on the target operator to obtain a preprocessed mixed integer programming equation; The preprocessed mixed integer programming equation is solved, and the obtained solution is used as the processing result of the task. The device according to claim 14, wherein the information includes at least one of the following: dimensions of variables of the mixed integer programming equation, sparsity of the constraint matrix of the mixed integer programming equation, the constraint matrix The blocking information or the rows in the constraint matrix that have a computational relationship with the variable, wherein the blocking information is used to indicate a submatrix whose elements are zero contained in the constraint matrix. The device according to claim 14 or 15, characterized in that the processor is used for: In the equation resource pool, determine a preset equation that matches the information related to the structure of the mixed integer programming equation; In the operator resource pool, the operator corresponding to the preprocessing of the preset equation is determined as the target operator. A computer storage medium, characterized in that the computer storage medium stores a computer program, which when executed by a computer causes the computer to implement the method described in any one of claims 1 to 12. A computer program product, characterized in that the computer program product stores instructions, which when executed by a computer, cause the computer to implement the method described in any one of claims 1 to 12.