CN107818365A - Merge test function optimization method, the device of cuckoo searching algorithm and wolf pack algorithm - Google Patents

Abstract

The present invention discloses a kind of test function optimization method and device for merging cuckoo searching algorithm and wolf pack algorithm, is related to function optimizing field.Wherein, methods described is assessed by using four selfs variable value weighting corresponding to preceding m caused by iteration less 4th test function values difference and produces new independent variable numerical value, and the new independent variable numerical value is added in optimal independent variable sequence of values, improve the convergence rate and precision of test function optimizing.

Description

Test function optimization method and device integrating cuckoo search algorithm and wolf pack algorithm

technical field

The present invention relates to the field of function optimization, in particular to a test function optimization method and device combining cuckoo search algorithm and wolf pack algorithm.

Background technique

Cuckoo Search Algorithm (CS) is a new meta-heuristic search algorithm jointly developed by Xin-she Yang et al. through observation of magical natural phenomena and artificial processing. The specific steps of the cuckoo search algorithm are as follows:

1. Perform initialization settings

Randomly generate the positions of N bird nests Then bring the positions of the N bird's nests to the test function for a selection test. After testing, the position of the best bird's nest can be selected and saved for the next generation.

2. Bird nest location search

Based on Levi's flight, start to search for the position of the next-generation bird's nest through the following position update formula, so as to obtain a new set of bird's nest positions, and bring them into the test function again for testing experiments, keeping up with the bird's nest obtained from the previous generation After comparing the nest positions, select the optimal bird nest position and enter the next step.

3. Selection of bird nest location

Comparing the probability p _a =0.25 that the host finds a foreign egg with the random number r∈[0,1] that obeys the uniform distribution, if r>p _a , then we can randomly change value; otherwise, it is fixed. Then use the test function to test the changed bird’s nest position, compare the tested bird’s nest position with the optimal bird’s nest position of the previous generation, and record the bird’s nest position with the best test value Finally, select the nest position with the best test value

4. Accuracy or iterative judgment

calculate Whether the target accuracy is reached or the iteration termination condition is met. If the requirements are met, then It is the global optimal solution gb, if it does not meet, it should be kept To the next generation, return to step 2, and then update iteratively.

The wolf pack algorithm mainly simulates the leadership hierarchy and hunting mechanism of wolves in nature, and divides wolves into four types by simulating the leadership hierarchy of wolves, as shown in Figure 1. α, β, δ are regarded as the top three wolves with the best performance in the wolf pack, and they guide other wolves (ω) towards the best area in the search space. In the whole iterative search process, α, β, δ three kinds of wolves are used to predict and evaluate the possible position of the prey.

According to the hunting mechanism of wolves, the hunting process is divided into three stages: searching for prey, surrounding prey, and attacking prey, as shown in Figure 2. In the wolf pack algorithm, the position of the three solutions is used to comprehensively evaluate the position of the optimal solution, which can greatly reduce the probability of the algorithm falling into a local optimum.

However, the inventors of the present application have found that when using the existing cuckoo search algorithm to optimize the test function, there are disadvantages of slow convergence speed, weak local search ability and premature convergence.

Contents of the invention

The purpose of the present invention is to provide a test function optimization method and device that integrates the cuckoo search algorithm and the wolf pack algorithm. Among them, the technical problem to be solved by the method is: how to solve the problems of slow convergence speed, weak local search ability and premature convergence when the existing cuckoo search algorithm optimizes the test function.

In order to achieve the above object, the present invention provides a test function optimization method that combines the cuckoo search algorithm and the wolf pack algorithm. The method comprises the steps of:

S1. Randomly generate n first independent variable values according to the independent variable value range of the test function, and respectively substitute the n first independent variable values into the test function to obtain the corresponding value of each first independent variable value The first test function value of ;

S2. According to the Levy flight mechanism, the values of the n first independent variables are respectively updated to obtain n corresponding values of the second independent variables, and the values of the n second independent variables are respectively substituted into the test function, obtaining a second test function value corresponding to each second independent variable value;

S3. Comparing the first test function value corresponding to each first independent variable value with the second test function value corresponding to the second independent variable value, and retaining the independent variable value corresponding to the smaller function value to obtain nth number Three independent variable values;

S4. Respectively update the values of the n third independent variables according to the preset discovery probability to obtain n values of the fourth independent variable;

S5. Determine whether the number of iterations reaches the preset number of iterations; if not, execute step S6; if yes, execute step S7;

S6. Select m fourth independent variable values from the n fourth independent variable values according to the fourth test function values respectively corresponding to the n fourth independent variable values, and select m fourth independent variable values for the m fourth independent variable values performing weighted evaluation to obtain a new independent variable value, and assigning the m fourth independent variable values and the new independent variable values to number positions respectively, and including the assigned m fourth independent variable values The variable values and the n independent variable values of the new independent variable values are used as n first independent variable values, and return to step S1, wherein n and m are constants respectively;

S7. Outputting the smallest fourth test function value and the fourth independent variable value corresponding to the smallest fourth test function value.

Optionally, according to the Levy flight mechanism, the values of the n first independent variables are respectively updated to obtain n corresponding values of the second independent variables, including:

The values of the n first independent variables are respectively updated according to the following formula one:

Among them, t represents the current number of iterations, x _i ^(t+1) represents the value of the i-th second independent variable after the update, x _i ^t represents the current value of the i-th first independent variable, and α represents the value of the cuckoo search path control parameter, Represents the point product, and levy(λ) represents the random search path of cuckoo flight.

Optionally, step S4 further includes:

Use the random number r∈[0,1] that obeys the uniform distribution as the discovery probability of the value of the third independent variable;

Comparing the random number r with a preset discovery probability to obtain a comparison result;

When it is judged according to the comparison result that the random number r is greater than the preset issuance probability, randomly update the value of the third independent variable;

Substituting the updated third independent variable value into the test function to obtain a test function value corresponding to the updated third independent variable value;

Comparing the third test function value corresponding to the third independent variable value with the updated test function value corresponding to the third independent variable value, retaining the independent variable value corresponding to the smaller function value, and obtaining the fourth independent variable value .

Optionally, the selecting m fourth independent variable values from the n fourth independent variable values according to the fourth test function values respectively corresponding to the n fourth independent variable values further includes:

Arranging the fourth test function values respectively corresponding to the n fourth independent variable values in ascending order to obtain an arrangement result;

Selecting the fourth independent variable values corresponding to the first m smaller fourth test function values according to the arrangement result, so as to select m fourth independent variable values from the n fourth independent variable values,

Wherein, m is a constant less than or equal to n.

Optionally, the weighted evaluation of the m fourth independent variable values to obtain a new independent variable value further includes:

According to the weighted evaluation of the following formula two, the new independent variable value is obtained:

Among them, x _new represents the value of the new independent variable, m represents the constant, x ₁ , x ₂ ,..., x _m represent the value of the fourth independent variable respectively, and x ₁ , x ₂ ,..., x _m respectively correspond to The value of the fourth test function of increases sequentially, and λ ₁ , λ ₂ ,...,λ _m respectively represent the proportions of the fourth independent variable values x ₁ , x ₂ ,...,x _m in the weighted evaluation.

Optionally, the assigning the m fourth independent variable values and the new independent variable values to numbered positions respectively, further includes:

assigning the fourth independent variable value x _k to numbering position h _k-1 +1 to numbering position h _{k respectively} ;

assigning said new independent variable values to numbered position h _k +1 to numbered position h respectively,

Among them, k, h _k-1 , h _k and h all represent constants.

Correspondingly, the present invention also provides a test function optimization device that integrates the cuckoo search algorithm and the wolf pack algorithm. The devices include:

A random generating unit, configured to randomly generate n first independent variable values according to the value range of the independent variables of the test function, and substitute the n first independent variable values into the test function respectively to obtain each first The value of the first test function corresponding to the value of the independent variable;

The first update unit is used to update the n first independent variable values according to the Levi's flight mechanism to obtain n corresponding second independent variable values, and respectively substitute the n second independent variable values into The test function obtains a second test function value corresponding to each second independent variable value;

The comparison unit is used to compare the first test function value corresponding to each first independent variable value with the second test function value corresponding to the second independent variable value, and retain the independent variable value corresponding to the smaller function value to obtain n values of the third independent variable;

The second updating unit is configured to respectively update the values of the n third independent variables according to a preset discovery probability to obtain n values of the fourth independent variables;

A judging unit, configured to judge whether the number of iterations reaches a preset number of iterations, and if not, select from the n fourth independent variable values according to the fourth test function values respectively corresponding to the n fourth independent variable values m fourth independent variable values, weighted evaluation of the m fourth independent variable values to obtain a new independent variable value, and the m fourth independent variable values and the new independent variable value respectively Assign to the numbered position, and use the n independent variable values including the assigned m fourth independent variable values and the new independent variable values as n first independent variable values, and return to step S1, wherein, n and m are constants respectively, if yes, output the smallest fourth test function value and the fourth independent variable value corresponding to the smallest fourth test function value.

Optionally, the judging unit is specifically configured to:

Arranging the fourth test function values respectively corresponding to the n fourth independent variable values in ascending order to obtain an arrangement result;

Selecting the fourth independent variable values corresponding to the first m smaller fourth test function values according to the arrangement result, so as to select m fourth independent variable values from the n fourth independent variable values,

Wherein, m is a constant less than or equal to n.

Optionally, the judging unit is also used for:

According to the weighted evaluation of the following formula two, the new independent variable value is obtained:

Among them, x _new represents the value of the new independent variable, m represents the constant, x ₁ , x ₂ ,..., x _m represent the value of the fourth independent variable respectively, and x ₁ , x ₂ ,..., x _m respectively correspond to The value of the fourth test function of increases sequentially, and λ ₁ , λ ₂ ,...,λ _m respectively represent the proportions of the fourth independent variable values x ₁ , x ₂ ,...,x _m in the weighted evaluation.

Optionally, the judging unit is also used for:

assigning the fourth independent variable value x _k to numbering position h _k-1 +1 to numbering position h _{k respectively} ;

assigning said new independent variable values to numbered position h _k +1 to numbered position h respectively,

Among them, k, h _k-1 , h _k and h all represent constants.

It can be seen from the above technical solution that, in the case where it is judged that the number of iterations has not reached the preset number of iterations, m number of fourth independent variable values are selected from n fourth independent variable values according to the fourth test function values respectively corresponding to n fourth independent variable values. Four independent variable values, perform weighted evaluation on m fourth independent variable values to obtain a new independent variable value, and assign the m fourth independent variable values and the new independent variable value to the numbered positions respectively, and include the assignment The last m fourth independent variable values and n independent variable values of the new independent variable values are used as the n first independent variable values for the next iteration, which not only improves the convergence speed and accuracy of the test function optimization, but also reduces The probability of the test function falling into a local optimal solution is determined.

Description of drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings that are required in the description of the embodiments or the prior art. Apparently, the drawings in the following description are only some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

Fig. 1 is the schematic diagram of the leadership hierarchy of simulating wolf pack in the prior art;

Fig. 2 is a schematic diagram of utilizing the wolf pack algorithm to find an optimal solution in the prior art;

Fig. 3 is the flow chart of the test function optimization method of fusion cuckoo search algorithm and wolf pack algorithm provided by an embodiment of the present invention;

Fig. 4 is the three-dimensional surface diagram of the Rastrigrin function provided by an embodiment of the present invention;

Fig. 5 is a curve diagram of a single convergence of the Rastrigrin function provided by an embodiment of the present invention;

Fig. 6 is a curve diagram of the mean convergence of the Rastrigrin function provided by an embodiment of the present invention;

Fig. 7 is a schematic structural diagram of a test function optimization device integrating a cuckoo search algorithm and a wolf pack algorithm provided by an embodiment of the present invention.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

The existing cuckoo search algorithm uses the Levi's flight mechanism to conduct a highly random search for the bird's nest, and does not make full use of the found better bird's nest position during the search process. Therefore, when using the existing cuckoo search algorithm to test When the function is optimized, it will affect the convergence speed and accuracy of the optimization. Based on this, the present invention provides a test function optimization method combining the cuckoo search algorithm and the wolf pack algorithm.

Fig. 3 is a flow chart of a test function optimization method integrating cuckoo search algorithm and wolf pack algorithm provided by an embodiment of the present invention. Optimizing, for example, finds the minimum value of the test function and the value of the independent variable corresponding to the minimum value, or finds the maximum value of the test function and the value of the independent variable corresponding to the maximum value. As shown in Figure 3, the test function optimization method that integrates the cuckoo search algorithm and the wolf pack algorithm provided by an embodiment of the present invention includes the following steps:

In step S1, according to the independent variable value range of the test function, n first independent variable values are randomly generated, and the n first independent variable values are respectively substituted into the test function to obtain each first independent variable value The first test function value corresponding to the variable value.

Among them, before step S1, it is also necessary to set a specific test function, a value range of an independent variable, and a size of a constant n. For example, the test function may be a test function Rastrigrin, the value range of the independent variable is x∈[-5.12,5.12], the global optimal solution of the test function is 0, distributed at (0,0), The constant n is 30. Specifically, the test function Rastrigrin is:

Wherein, the dimension of the function is d=30, and the value of the first independent variable represents the value of the independent variable randomly generated according to the value range of the independent variable of the test function.

Next, in step S2, the values of the n first independent variables are respectively updated according to the Levi's flight mechanism to obtain n corresponding values of the second independent variables, and the values of the n second independent variables are respectively substituted into The test function obtains a second test function value corresponding to each second independent variable value.

Specifically, according to the Levy flight mechanism, the values of the n first independent variables are respectively updated to obtain n corresponding second independent variable values, including:

The values of the n first independent variables are respectively updated according to the following formula one:

Among them, t represents the current number of iterations, x _i ^(t+1) represents the value of the i-th second independent variable after the update, x _i ^t represents the current value of the i-th first independent variable, and α represents the value of the cuckoo search path control parameter, Represents the point product, and levy(λ) represents the random search path of cuckoo flight.

Wherein, the second independent variable value represents the independent variable value obtained by updating the randomly generated independent variable value by the Levy flight mechanism.

Next, in step S3, the first test function value corresponding to each first independent variable value is compared with the second test function value corresponding to the second independent variable value, and the independent variable corresponding to the smaller function value is retained Value, to get n values of the third independent variable.

Specifically, when the first test function value corresponding to the first independent variable value is smaller than the second test function value corresponding to the second independent variable value corresponding to the first independent variable value, keep the first independent variable value and discard The value of the second independent variable corresponding to the value of the first independent variable; the value of the first test function corresponding to the value of the first independent variable is greater than the value of the second test function corresponding to the value of the second independent variable corresponding to the value of the first independent variable In this case, the value of the second independent variable corresponding to the value of the first independent variable is retained, and the value of the first independent variable is discarded; the value of the first test function corresponding to the value of the first independent variable is equal to the value of the first independent variable corresponding to the value of the first independent variable In the case of the second test function value of two independent variable values, either one of the first independent variable value and the second independent variable value corresponding to the first independent variable value is retained.

Wherein, the value of the third independent variable represents the value of the independent variable corresponding to the test function value corresponding to the value of the first independent variable and the value of the second independent variable which is smaller in the value of the test function.

Then, in step S4, the values of the n third independent variables are respectively updated according to the preset discovery probability to obtain n values of the fourth independent variable.

Specifically, this step further includes: using a random number r∈[0,1] subject to uniform distribution as the discovery probability of the value of the third independent variable; comparing the random number r with the preset discovery probability to obtain a comparison result ; In the case of judging that the random number r is greater than the preset issuance probability according to the comparison result, randomly updating the value of the third independent variable; substituting the updated value of the third independent variable into the test function, Obtain the test function value corresponding to the updated third independent variable value; compare the third test function value corresponding to the third independent variable value with the test function value corresponding to the updated third independent variable value, and retain the function value The value of the independent variable corresponding to the smaller one, to obtain the value of the fourth independent variable.

Wherein, the random updating of the value of the third independent variable includes: updating the value of the third independent variable by adopting the bird's nest location random preference walk update formula in the prior art. Wherein, when the third test function value corresponding to the value of the third independent variable is smaller than the value of the test function corresponding to the updated value of the third independent variable, the value of the third independent variable is retained, and the updated value of the third independent variable is discarded; When the value of the third test function corresponding to the value of the third independent variable is greater than the value of the test function corresponding to the updated value of the third independent variable, keep the updated value of the third independent variable and discard the value of the third independent variable; When the third test function value corresponding to the three independent variable values is equal to the test function value corresponding to the updated third independent variable value, any one of the third independent variable value and the updated third independent variable value is retained.

Wherein, the value of the fourth independent variable represents the value of the independent variable obtained after updating the value of the third independent variable according to the preset discovery probability.

Next, in step S5, it is judged whether the number of iterations reaches a preset number of iterations.

Next, in step S6, when it is judged that the number of iterations has not reached the preset number of iterations, according to the fourth test function values respectively corresponding to the values of the n fourth independent variables, the n fourth independent variables Selecting m fourth independent variable values from the values, performing weighted evaluation on the m fourth independent variable values to obtain a new independent variable value, and combining the m fourth independent variable values with the new independent variable value Assign variable values to numbered positions respectively, and use n independent variable values including the assigned m fourth independent variable values and the new independent variable values as n first independent variable values, and return to step S1 , where n and m are constants respectively.

Finally, in step S7, when it is judged that the number of iterations reaches the preset number of iterations, the smallest fourth test function value and the fourth argument value corresponding to the smallest fourth test function value are output.

Specifically, the selecting m fourth independent variable values from the n fourth independent variable values according to the fourth test function values respectively corresponding to the n fourth independent variable values further includes: The fourth test function values corresponding to the fourth independent variable values are arranged in ascending order to obtain the arrangement result; according to the arrangement result, select the first m smaller fourth test function values corresponding to the fourth self respectively. variable values, so as to select m fourth independent variable values from the n fourth independent variable values, wherein m is a constant less than or equal to n.

For example, when n=30 and m=3, arrange the fourth test function values corresponding to the 30 fourth independent variable values in ascending order to obtain the arrangement result; select the first three according to the arrangement result The fourth independent variable values x ₁ , x ₂ , x ₃ corresponding to the smaller fourth test function values respectively, in other words, the fourth independent variable values x ₁ , x ₂ , x ₃ respectively correspond to the fourth test function values Increase in turn, so as to realize the selection of 3 fourth independent variable values from 30 fourth independent variable values.

Specifically, the weighted evaluation of the m fourth independent variable values to obtain a new independent variable value further includes:

According to the weighted evaluation of the following formula two, the new independent variable value is obtained:

Among them, x _new represents the value of the new independent variable, m represents the constant, x ₁ , x ₂ ,..., x _m represent the value of the fourth independent variable respectively, and x ₁ , x ₂ ,..., x _m respectively correspond to The value of the fourth test function of increases sequentially, and λ ₁ , λ ₂ ,...,λ _m respectively represent the proportions of the fourth independent variable values x ₁ , x ₂ ,...,x _m in the weighted evaluation.

For example, when m=3, the proportions of the fourth independent variable values x ₁ , x ₂ , and x ₃ in the weighted evaluation are λ ₁ =1.2, λ ₂ =1, and λ ₃ =0.8 respectively, so that The new argument value.

Specifically, assigning the m fourth independent variable values and the new independent variable values to numbered positions respectively, further includes: respectively assigning the fourth independent variable values x _k to numbered positions h _k-1 + 1 to number position h _k ; assign the new independent variable value to number position h _k +1 to number position h respectively, wherein k, h _k-1 , h _k and h all represent constants.

For example, when m=3, h=8, h ₁ =2, h ₂ =4, h ₃ =6, the value x ₁ of the fourth argument is assigned to the number position 1 and the number position 2 respectively, and the fourth argument The variable value x ₂ is assigned to number position 3 and number position 4 respectively, the fourth independent variable value x ₃ is assigned to number position 5 and number position 6 respectively, and the new independent variable value is assigned to number position 7 and number position 8 respectively. Specifically, the 30 fourth independent variable values correspond to a numbered position respectively. After obtaining the fourth independent variable values x ₁ , x ₂ , x ₃ and the new independent variable values, the fourth independent variable values corresponding to the original numbered positions 1-8 The argument value is discarded, leaving numbered positions 1-8 empty. Then, assign the fourth independent variable value x ₁ , x ₂ , x ₃ and the new independent variable value to numbered positions 1-8 according to the above method.

When multiple independent variable values are selected, the local search in multiple optimal regions can be strengthened at the same time. Adjusting the number h of numbered positions and the number m of optimal independent variable values can balance the capabilities of global search and local search.

Taking the test function Rastrigrin as an example, combining the existing cuckoo search algorithm and the improved cuckoo search algorithm to evaluate the performance advantages of the test function optimization method integrating the cuckoo search algorithm and the wolf pack algorithm provided by the present invention. Among them, the three-dimensional surface diagram of the Rastrigrin function is shown in Figure 4.

When using the existing cuckoo search algorithm to optimize the test function, the algorithm runs independently 50 times, and the number of iterations of each independent operation is 100. When using the improved cuckoo search algorithm to optimize the test function, the algorithm runs independently 50 times, the number of iterations of each independent operation is 100, when using the test function optimization method of the fusion cuckoo search algorithm and wolf pack algorithm provided by the present invention to optimize the test function, the algorithm runs 50 times independently, each time independently The number of iterations run is 100. Among them, the difference between the improved cuckoo search algorithm and the method provided by the present invention is that the improved cuckoo search algorithm assigns the selected independent variable values to the numbered positions respectively, but does not generate new independent variable values.

Fig. 5 is a graph of a single convergence of the Rastrigrin function provided by an embodiment of the present invention. Fig. 5 shows the convergence curve of the test function Rastrigrin in one of the 50 independent runs, where CS represents the existing cuckoo search algorithm, ICS represents the improved cuckoo search algorithm, and new-ICS represents the current The invention provides a test function optimization method that combines the cuckoo search algorithm and the wolf pack algorithm. It can be seen from the figure that, compared with the existing cuckoo search algorithm and the improved cuckoo search algorithm, the method provided by the present invention has the advantage of fast convergence speed.

Fig. 6 is a graph showing the average convergence of the Rastrigrin function provided by an embodiment of the present invention. Figure 6 shows a graph of the convergence of the average of the searched minima of the test function over 50 independent runs. It can be seen from the figure that, compared with the existing cuckoo search algorithm and the improved cuckoo search algorithm, the method provided by the present invention has the advantage of fast convergence speed.

When the number of iterations is 500, the optimization accuracy of the three methods is compared as follows:

CS: best=114.387, mean=150.389, max=196.027, std=16.7872,

ICS: best=61.7044, mean=93.1059, max=136.066, std=17.8358,

new-ICS: best=27.1544, mean=59.3567, max=141.707, std=20.5961

Among them, best represents the optimal value, mean represents the average value, max represents the maximum value, and Std represents the standard value. It can be seen from the comparison that the optimization accuracy of the method provided by the present invention is greater than that of the existing cuckoo search algorithm and the improved cuckoo search algorithm.

In this embodiment, when it is judged that the number of iterations has not reached the preset number of iterations, m fourth independent variables are selected from the n fourth independent variable values according to the fourth test function values corresponding to the n fourth independent variable values. Numerical value, carry out weighted evaluation on m fourth independent variable values to obtain a new independent variable value, and assign m fourth independent variable values and new independent variable values to number positions respectively, and include the assigned m The fourth independent variable value and n independent variable values of the new independent variable value are used as n first independent variable values for the next iteration, which not only improves the convergence speed and accuracy of the test function optimization, but also reduces the test function The probability of finding a local optimal solution.

For the method embodiment, for the sake of simple description, it is expressed as a series of action combinations, but those skilled in the art should know that the embodiment of the present invention is not limited by the described action order, because according to the embodiment of the present invention , certain steps may be performed in other order or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification belong to preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

Fig. 7 is a schematic structural diagram of a test function optimization device integrating a cuckoo search algorithm and a wolf pack algorithm provided by an embodiment of the present invention. As shown in FIG. 7 , the test function optimization device that integrates the cuckoo search algorithm and the wolf pack algorithm provided by an embodiment of the present invention includes:

The random generating unit 201 is configured to randomly generate n first independent variable values according to the value range of the independent variables of the test function, and substitute the n first independent variable values into the test function to obtain each first independent variable value. a first test function value corresponding to an independent variable value;

The first updating unit 202 is configured to update the values of the n first independent variables according to the Levi's flight mechanism to obtain n corresponding values of the second independent variables, and respectively update the values of the n second independent variables Substituting into the test function to obtain a second test function value corresponding to each second independent variable value;

The comparison unit 203 is configured to compare the first test function value corresponding to each first argument value with the second test function value corresponding to the second argument value, and retain the argument value corresponding to the smaller function value, Obtain n values of the third independent variable;

The second updating unit 204 is configured to respectively update the values of the n third independent variables according to a preset discovery probability to obtain n values of the fourth independent variables;

The judging unit 205 is configured to judge whether the number of iterations has reached the preset number of iterations, and if not, select from the n fourth independent variable values according to the fourth test function values respectively corresponding to the n fourth independent variable values Selecting m fourth independent variable values, performing weighted evaluation on the m fourth independent variable values to obtain a new independent variable value, and combining the m fourth independent variable values with the new independent variable value respectively assigning to numbered positions, and using n independent variable values including the assigned m fourth independent variable values and the new independent variable values as n first independent variable values, and returning to step S1, wherein , n and m are constants respectively, if so, output the smallest fourth test function value and the fourth independent variable value corresponding to the smallest fourth test function value.

In an optional implementation manner of the present invention, the judging unit 205 is specifically configured to:

Arranging the fourth test function values respectively corresponding to the n fourth independent variable values in ascending order to obtain an arrangement result;

Selecting the fourth independent variable values corresponding to the first m smaller fourth test function values according to the arrangement result, so as to select m fourth independent variable values from the n fourth independent variable values,

Wherein, m is a constant less than or equal to n.

In an optional implementation manner of the present invention, the judging unit 205 is further configured to:

According to the weighted evaluation of the following formula two, the new independent variable value is obtained:

Among them, x _new represents the value of the new independent variable, m represents the constant, x ₁ , x ₂ ,..., x _m represent the value of the fourth independent variable respectively, and x ₁ , x ₂ ,..., x _m respectively correspond to The value of the fourth test function of increases sequentially, and λ ₁ , λ ₂ ,...,λ _m respectively represent the proportions of the fourth independent variable values x ₁ , x ₂ ,...,x _m in the weighted evaluation.

In an optional implementation manner of the present invention, the judging unit 205 is further configured to:

assigning the fourth independent variable value x _k to numbering position h _k-1 +1 to numbering position h _{k respectively} ;

assigning said new independent variable values to numbered position h _k +1 to numbered position h respectively,

Among them, k, h _k-1 , h _k and h all represent constants.

For the specific details involved in the test function optimization device integrating the cuckoo search algorithm and the wolf pack algorithm provided by an embodiment of the present invention, the test function of the fusion cuckoo search algorithm and the wolf pack algorithm provided by an embodiment of the present invention The optimization method has been described in detail, and will not be repeated here.

It should be noted that among the various components of the system of the present invention, the components are logically divided according to the functions to be realized, but the present invention is not limited thereto, and each component can be re-divided as required Or combined, for example, some components can be combined into a single component, or some components can be further broken down into more subcomponents.

The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor (DSP) may be used in practice to implement some or all functions of some or all components in the system according to the embodiments of the present invention. The present invention can also be implemented as an apparatus or an apparatus program (for example, a computer program and a computer program product) for performing a part or all of the methods described herein. Such a program for realizing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet site, or provided on a carrier signal, or provided in any other form.

It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means can be embodied by one and the same item of hardware. The use of the words first, second, and third, etc. does not indicate any order. These words can be interpreted as names.

The above embodiments are only suitable for illustrating the present invention, rather than limiting the present invention. Those of ordinary skill in the relevant technical fields can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, all Equivalent technical solutions also belong to the category of the present invention, and the scope of patent protection of the present invention should be defined by the claims.

Claims (10)

1. A kind of 1. test function optimization method for merging cuckoo searching algorithm and wolf pack algorithm, it is characterised in that methods described Comprise the following steps：

   S1, generate n the first independent variable numerical value at random according to the independent variable span of the test function, and by the n First independent variable numerical value substitutes into the test function respectively, obtains the first test function corresponding to every one first independent variable numerical value Value；

   S2, fly mechanics are tieed up according to Lay the n the first independent variable numerical value are updated respectively, obtained second corresponding to n Independent variable numerical value, and the n the second independent variable numerical value are substituted into the test function respectively, obtain every one second independent variable number Second test function value corresponding to value；

   S3, the second survey by the first test function value corresponding to every one first independent variable numerical value with corresponding second independent variable numerical value Trial function value is compared, and independent variable numerical value corresponding to reservation functional value smaller, obtains n the 3rd independent variable numerical value；

   S4, according to default probability of detection the n the 3rd independent variable numerical value are updated respectively, obtain n four selfs and become Numerical quantity；

   S5, judge whether iterations reaches default iterations；If it is not, then perform step S6；If so, then perform step S7；

   S6, according to the n four selfs variable values respectively corresponding to the 4th test function value from the n four selfs variable numbers M four selfs variable values are selected in value, assessments is weighted to the m four selfs variable values, obtain one it is new certainly Variable value, and the m four selfs variable values and the new independent variable numerical value are respectively allocated to numbered positions, and will Comprising the m four selfs variable values after distribution and n independent variable numerical value of the new independent variable numerical value as n individual the One independent variable numerical value, and return to step S1, wherein, n and m are respectively constant；

   Four selfs variable value corresponding to 4th test function value of the 4th minimum test function value of S7, output and minimum.
2. 2. the test function optimization method of fusion cuckoo searching algorithm according to claim 1 and wolf pack algorithm, it is special Sign is, described that the n the first independent variable numerical value are updated respectively according to Lay dimension fly mechanics, obtains corresponding to n Second independent variable numerical value, including：

   The n the first independent variable numerical value are updated according to below equation one respectively：

   Wherein, t represents current iteration number, x_i ^(t+1)Represent i-th of second independent variable numerical value after renewal, x_i ^tRepresent currently I-th of first independent variable numerical value, α represent the control parameter of cuckoo searching route,Dot-product is represented, levy (λ) represents cloth Paddy bird flying random search path.
3. 3. the test function optimization method of fusion cuckoo searching algorithm according to claim 1 and wolf pack algorithm, it is special Sign is, step S4, further comprises：

   Using probability of detection of the equally distributed random number r ∈ [0,1] of obedience as the 3rd independent variable numerical value；

   By the random number r compared with default probability of detection, comparative result is obtained；

   In the case where judging that the random number r is more than the default distribution probability according to the comparative result, randomly update The 3rd independent variable numerical value；

   By the 3rd independent variable numerical value substitution test function after renewal, corresponding to the 3rd independent variable numerical value after being updated Test function value；

   By the survey corresponding with the 3rd independent variable numerical value after renewal of the 3rd test function value corresponding to the 3rd independent variable numerical value Trial function value is compared, and independent variable numerical value corresponding to reservation functional value smaller, obtains four selfs variable value.
4. 4. the test function optimization method of fusion cuckoo searching algorithm according to claim 1 and wolf pack algorithm, it is special Sign is that the 4th test function value according to corresponding to the n four selfs variable values difference becomes from the n four selfs M four selfs variable values are selected in numerical quantity, are further comprised：

   By the n four selfs variable values, corresponding 4th test function value is arranged by order from small to large respectively, Obtain rank results；

   Four selfs variable value corresponding to m less 4th test function values difference before being selected according to the rank results, from And realize and m four selfs variable values are selected from the n four selfs variable values,

   Wherein, m is the constant less than or equal to n.
5. 5. the test function optimization method of fusion cuckoo searching algorithm according to claim 1 and wolf pack algorithm, it is special Sign is, described to be weighted assessment to the m four selfs variable values, obtains a new independent variable numerical value, further Including：

   Assessed according to the weighting of below equation two and obtain the new independent variable numerical value：

   Wherein, x_newNew independent variable numerical value is represented, m represents constant, x₁,x₂,...,x_mFour selfs variable value is represented respectively, and x₁,x₂,...,x_mThe 4th test function value increases successively corresponding to respectively, λ₁,λ₂,...,λ_mFour selfs variable value is represented respectively x₁,x₂,...,x_mThe shared proportion in weighting is assessed.
6. 6. the test function optimization method of fusion cuckoo searching algorithm according to claim 1 and wolf pack algorithm, it is special Sign is, described that the m four selfs variable values and the new independent variable numerical value are respectively allocated into numbered positions, enters one Step includes：

   By four selfs variable value x_kIt is separately dispensed into numbered positions h_k-1+ 1 to numbered positions h_k；

   The new independent variable numerical value is separately dispensed into numbered positions h_k+ 1 to numbered positions_h, wherein, k, h_k-1、h_kAnd h is equal Represent constant.
7. A kind of 7. test function optimization device for merging cuckoo searching algorithm and wolf pack algorithm, it is characterised in that described device Including：

   Random generation unit, for generating n the first independent variable numbers at random according to the independent variable span of the test function Value, and the n the first independent variable numerical value are substituted into the test function respectively, obtain corresponding to every one first independent variable numerical value First test function value；

   First updating block, the n the first independent variable numerical value are updated respectively for tieing up fly mechanics according to Lay, obtained Second independent variable numerical value corresponding to n, and the n the second independent variable numerical value are substituted into the test function respectively, obtain every Second test function value corresponding to one second independent variable numerical value；

   Comparing unit, for by the first test function value corresponding to every one first independent variable numerical value and corresponding second independent variable number Second test function value of value is compared, and independent variable numerical value corresponding to reservation functional value smaller, obtains n the 3rd independents variable Numerical value；

   Second updating block, for being updated respectively to the n the 3rd independent variable numerical value according to default probability of detection, obtain To n four selfs variable values；

   Judging unit, for judging whether iterations reaches default iterations, if it is not, then according to the n four selfs The 4th test function value selects m four selfs variable numbers from the n four selfs variable values corresponding to variable value difference Value, assessment is weighted to the m four selfs variable values, obtains a new independent variable numerical value, and by the m the 4th Independent variable numerical value and the new independent variable numerical value are respectively allocated to numbered positions, and will include the m the 4th after distribution N independent variable numerical value of independent variable numerical value and the new independent variable numerical value is as n the first independent variable numerical value, and return to step S1, wherein, n and m are respectively constant, if so, then exporting the 4th test function value of minimum and the 4th minimum test function value Corresponding four selfs variable value.
8. 8. the test function optimization device of fusion cuckoo searching algorithm according to claim 7 and wolf pack algorithm, it is special Sign is, the judging unit, is specifically used for：

   By the n four selfs variable values, corresponding 4th test function value is arranged by order from small to large respectively, Obtain rank results；

   Four selfs variable value corresponding to m less 4th test function values difference before being selected according to the rank results, from And realize and m four selfs variable values are selected from the n four selfs variable values,

   Wherein, m is the constant less than or equal to n.
9. 9. the test function optimization device of fusion cuckoo searching algorithm according to claim 7 and wolf pack algorithm, it is special Sign is, the judging unit, is additionally operable to：

   Assessed according to the weighting of below equation two and obtain the new independent variable numerical value：

   Wherein, x_newNew independent variable numerical value is represented, m represents constant, x₁,x₂,...,x_mFour selfs variable value is represented respectively, and x₁,x₂,...,x_mThe 4th test function value increases successively corresponding to respectively, λ₁,λ₂,...,λ_mFour selfs variable value is represented respectively x₁,x₂,...,x_mThe shared proportion in weighting is assessed.
10. 10. the test function optimization device of fusion cuckoo searching algorithm according to claim 7 and wolf pack algorithm, it is special Sign is, the judging unit, is additionally operable to：

    By four selfs variable value x_kIt is separately dispensed into numbered positions h_k-1+ 1 to numbered positions h_k；

    The new independent variable numerical value is separately dispensed into numbered positions h_k+ 1 to numbered positions h,

    Wherein, k, h_k-1、h_kAnd h represents constant.