CN111556508A - A Stackelberg Game Multi-operator Dynamic Spectrum Sharing Method for Large-scale IoT Access - Google Patents

Abstract

A multi-operator dynamic spectrum sharing method for a Stackelberg game oriented to large-scale IoT access, the multi-operator dynamic spectrum sharing method comprises the following steps: 1) pre-establishing a benefit function of a spectrum provider and a network operator; 2) With the spectrum provider as the leader and the network operator as the follower, a dynamic Stackelberg game is established; 3) Through the game, an equilibrium is finally achieved, multi-operator spectrum sharing is achieved, and the benefits of the leader and the follower are maximized. Maximize the number of IoT devices entering followers; take the spectrum provider as the leader and the network operator as the follower, establish a dynamic Stackelberg game and operate under the dynamic spectrum sharing mechanism; through the game, a balance is finally reached, and the leader and the follower are realized. This maximizes the follower's revenue while maximizing the number of cellular users and IoT devices that can access the follower.

Description

A Stackelberg Game Multi-operator Dynamic Spectrum Sharing Method for Large-scale IoT Access

technical field

The invention belongs to the field of wireless communication and Internet of Things communication, and relates to a multi-operator dynamic spectrum sharing method oriented to large-scale IoT access based on Stackelberg game.

Background technique

With the rapid development of wireless communication technology, a large number of IoT devices will be connected to the Internet through 5G and other next-generation wireless communication networks. , see website, https://www.5gamericas.org/5g-the-future-of-iot/)”, the number of devices in the global IoT market will grow to 3 trillion by 2026. The wide application of IoT technology in the future will help people make great progress in smart city management, unmanned environmental monitoring and collaborative intelligent transportation systems, but it will also bring new challenges to the field of mobile wireless communications.

The main challenge at present is how wireless network operators can simultaneously accommodate the original high-speed cellular users and the rapidly growing large number of IoT devices in the limited spectrum resources while ensuring high communication service quality and low cost. Spectrum sharing technology is a promising solution, and the 3rd Generation Partnership Project (3GPP) organization proposed the idea of spectrum resource sharing in its release 14 Long Term Evolution (LTE) standard, which allows operators The spectrum resources of the wireless access network are shared with each other to improve the utilization rate of the spectrum resources of a single operator. While there is great potential for efficient spectrum allocation, such network sharing can greatly increase the complexity of wireless system implementation. In addition, the network sharing architecture proposed by 3GPP mainly focuses on high-speed downlink data transmission services for cellular users, while network sharing applied to a large number of IoT devices in the future will be dominated by a large number of uplink short-packet communications from terminal devices to base stations. Therefore, how to meet the bandwidth requirements of large-scale IoT devices while taking into account the original cellular users is still a problem.

After searching the existing literature, it was found that Y. Xiao et al. published a paper entitled "Multi-operator network sharing formassive IoT" in "IEEE Communications Magazine" in 2019. )" article. This article introduces a 3GPP-based multi-operator network sharing framework that supports coexistence of IoT devices and high-speed cellular service users, but does not consider how to dynamically share spectrum among multiple operators. It was also found after retrieval that N.N.Sapavath et al. published a paper entitled "Wirelessvirtualization architecture:Wireless networking for Internet of Things" in the "IEEE Internet of Things Journal (IEEE Transactions on the Internet of Things)" in 2019. Wireless Networks)" article. In this paper, the author proposes an iterative algorithm applied to a three-layer game model between wireless infrastructure providers, mobile virtual network operators and IoT devices, and proves the existence and uniqueness of game equilibrium points. However, how to design a simple and efficient algorithm to find the optimal sharing strategy in large-scale IoT scenarios is still a research focus.

To sum up, the existing problems are: (1) The traditional spectrum sharing technology is mainly used for downlink long packet communication from base stations to cellular mobile devices, while the future spectrum sharing will be mainly used for IoT terminal equipment Uplink short packet communication to the base station. (2) Some new spectrum sharing technologies do not achieve dynamic sharing among multiple wireless network operators, and the system lacks flexibility. (3) The computational complexity of the spectrum sharing algorithm is high, and it is difficult to implement in the IoT system with generally weak performance.

The significance of solving the above technical problems is: based on the current development of wireless communication technology and the improvement of the demand in the field of Internet of Things, a simpler and more efficient spectrum sharing method can realize the coexistence of large-scale Internet of Things devices and cellular users on the premise of ensuring the quality of service, and To jointly maximize the benefits of service providers and receivers, and promote the application and development of the Internet of Things.

SUMMARY OF THE INVENTION

Aiming at the problems existing in the prior art, the purpose of the present invention is to provide a multi-operator dynamic spectrum sharing method for Stackelberg game oriented to large-scale IoT access on the basis of ensuring service quality. It can greatly increase the number of cellular users and IoT devices that can be accessed under the condition of limited spectrum resources.

The object of the present invention is achieved in this way, a kind of Stackelberg game multi-operator dynamic spectrum sharing method for large-scale IoT access, comprising the following steps:

The multi-operator dynamic spectrum sharing method includes the following steps:

Step 1: Pre-establish the benefit function of spectrum providers and network operators;

Step 2: Establish a dynamic Stackelberg game with spectrum providers as leaders and network operators as followers;

Step 3: Through the game, the equilibrium is finally reached, the multi-operator spectrum sharing and the maximization of the income of the leader and the follower can be achieved, and the number of IoT devices connected to the follower can be maximized at the same time;

The spectrum provider is the manager of all available spectrum resources, and can formulate unit bandwidth prices and trade spectrum resources with the network operator in real time according to the purchase demand of the network operator; the network operator is the network service provider in a small area. Providers can update spectrum purchases in real time according to their own load conditions and unit bandwidth prices;

In the applicable network environment, cellular users and IoT devices exist at the same time, and the two follow the coexistence access rules when accessing the network: each cellular user uses orthogonal channels that do not interfere with each other, while IoT devices randomly reuse some sub-channels Transmission; after cellular users and IoT devices access the network operator, they will be charged only when the transmission rate exceeds a certain threshold; the network operator will give priority to ensuring the access and transmission of cellular users, and after all cellular users are satisfied will start to serve IoT devices;

赫兹；蜂窝用户和物联网设备适用不同的频谱分配方案，对于蜂窝用户，网络运营商以循环方式均等地调度频谱资源；对于物联网设备，每个设备从F个子信道中随机选择k个作为候选，并以p_m概率接入这k个子信道中的一个，这意味着有可能同时复用多个子信道进行传输；基于该分配方案，物联网设备选择子信道l(l∈{F})作为候选的概率为

接入该子信道的概率为p_lp_m；The multi-operator network environment is as follows: there is one spectrum provider, M network operators, the total available bandwidth owned by the spectrum provider is B ^tol Hz, and the spectrum purchased by the network operator m (m ∈ M) from the spectrum provider The bandwidth is B _m Hz; the available spectrum of network operator m is divided into F sub-channels, and the bandwidth of each sub-channel is

Hertz; different spectrum allocation schemes apply to cellular users and IoT devices. For cellular users, network operators schedule spectrum resources equally in a round-robin manner; for IoT devices, each device randomly selects k from F subchannels as candidates , and access one of the k sub-channels with p _m probability, which means that it is possible to multiplex multiple sub-channels for transmission at the same time; based on this allocation scheme, IoT devices select sub-channel l(l∈{F}) as the The candidate probability is

The probability of accessing the subchannel is p _l p _m ;

The benefit function U _w of spectrum providers and network operators is expressed as:

where r _p represents the spectrum fee charged from all network operators, C is the bandwidth license fee paid to the spectrum regulator, and P is the unit bandwidth price charged by the spectrum provider;

The benefit function U _m of the network operator can be expressed as:

和

分别为网络运营商m下蜂窝用户最大接入数量和物联网设备的最大共存接入数量，P_c和P_I分别表示网络运营商向每个蜂窝用户和物联网设备收取的接入费用；in

and

are the maximum access number of cellular users and the maximum coexisting access number of IoT devices under the network operator m, respectively, P _c and P _I represent the access fees charged by the network operator to each cellular user and IoT device;

In step 2: The dynamic Stackelberg game established with the spectrum provider as the leader and the network operator as the follower includes two sub-games:

a. Spectrum providers constantly adjust the unit bandwidth price P, and try to provide the most appropriate and favorable price to attract network operators to purchase more spectrum bandwidth, and ultimately maximize their own benefits. The sub-game of spectrum provider revenue maximization is expressed as :

s.t.P≥0

b. According to the price P announced by the spectrum provider and its own load situation, the network operator continuously submits appropriate bandwidth purchase requirements to maximize its own benefits, and since the benefits are positively related to the number of connected end customers, when the benefits are maximized Obtain the maximum number of cellular users and IoT device access; according to whether the bandwidth is sufficient to meet the needs of all cellular users, the subgame is divided into two cases:

①B _m ≤B _m,0 , insufficient bandwidth:

stB _m ≤B _m,0

②B _m ≥B _m,0 , the bandwidth is sufficient:

stB _m >B _m,0

The Stackelberg game is a classic economics game theory. It has certain advantages in solving the problem of resource scheduling. The participants in the game process are divided into two types: leaders and followers. The two compete asymmetrically for output. An output is determined first, and the follower observes the leader's decision before making a choice. The leader will also adjust his decision according to the follower's decision until an equilibrium is reached.

The dynamic spectrum sharing mechanism consists of two layers, namely, the per-bandwidth pricing control of the leader and the required bandwidth optimization of the followers. The leader's unit bandwidth pricing control: In order to allocate spectrum resources reasonably, spectrum providers conduct spectrum transactions with network operators by controlling and adjusting unit bandwidth prices. After collecting all feedback from network operators, spectrum providers will maximize their own interests. Pursue the optimization to update the pricing, and finally achieve the best unit block price that maximizes its own benefits; optimization of the required bandwidth of the followers: each network operator chooses the most purchased bandwidth according to its pursuit of maximizing benefits. It will comprehensively consider the bandwidth price and its own load, and feed back the required purchase amount to the spectrum provider.

The network operator's own load profile is derived from the number of cellular users and IoT devices within the service area requesting access. Within the service range of each network operator, cellular users and IoT devices can simultaneously use the licensed frequency band for uplink transmission, subject to coexistence access rules: each cellular user uses orthogonal channels that do not interfere with each other, while the Networked devices randomly reuse some sub-channels for transmission; after cellular users and IoT devices access network operators, they will be charged only when the transmission rate exceeds a certain threshold; network operators will give priority to ensuring cellular users’ access and Transmission, IoT devices will not be considered until all cellular users are satisfied. In the process of spectrum sharing, the network operator announces the access fee and rate threshold, and the subordinate cellular users and IoT devices decide whether to send an access request according to their own needs, and the network operator can obtain its own load according to the received request. .

The Stackelberg game running under the dynamic spectrum sharing mechanism in the step 2 needs to perform the following steps:

Step (2.1): Initialization is performed at the beginning of the game, and the spectrum provider broadcasts the preset initial spectrum pricing to all network operators.

Step (2.2): Each network operator tries to submit appropriate bandwidth purchase requirements according to the spectrum price announced by the spectrum provider and the load of its own cellular users and IoT devices, so as to maximize its own benefits.

Step (2.3): The spectrum provider trades spectrum with the network operator according to the bandwidth purchase demand submitted by the network operator, and calculates and adjusts the spectrum pricing in the next iteration according to a predetermined formula.

Step (2.4): When the game does not converge, the spectrum provider announces new pricing to all network operators at the beginning of the next iteration; repeat steps (2.2), (2.3) and (2.4) until the game reaches a specified At this time, the benefits of spectrum providers and network operators should be maximized, and the number of IoT devices that can access the network should also be maximized.

Beneficial effects of the present invention: Compared with the prior art, the present invention adopts a multi-operator dynamic spectrum sharing method based on Stackelberg game to realize joint dynamic spectrum sharing among spectrum providers, network operators, cellular users and IoT devices. decision making. Spectrum providers attract more bandwidth requests from network operators through appropriate prices, and network operators purchase the optimal amount of bandwidth according to their own load conditions and spectrum prices to maximize revenue, so that more users and devices can be connected. Compared with the traditional spectrum sharing scheme, the present invention greatly increases the accessible quantity of cellular users and IoT devices under the condition of limited spectrum resources. The invention can realize the coexistence of large-scale Internet of Things equipment and cellular users under the premise of ensuring service quality with a simple and efficient spectrum sharing method, and jointly maximize the benefits of the service provider and the receiver, and promote the use of existing communication conditions in the Internet of Things field. applications and development.

Description of drawings

FIG. 1 is a scene diagram of dynamic spectrum sharing implemented by the present invention.

Figure 2 is the dynamic Stackelberg flow chart used in the spectrum sharing implementation process.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention clearer, the following describes the embodiments of the present invention in detail with reference to the accompanying drawings: This embodiment is implemented on the premise of the technical solution of the present invention, and provides a detailed implementation manner and specific operating procedures. It should be understood that the specific examples described herein are only used to explain the present invention, but the protection scope of the present invention is not limited to the following embodiments.

Embodiment: This embodiment adopts the wireless communication scenario between the Internet of Things and cellular users as shown in FIG. 1, and proposes a Stackelberg game multi-operator dynamic spectrum sharing method for large-scale IoT access. In this scenario, the benefit functions U _w and U _m of the spectrum provider and the network operator need to be established first. The spectrum provider is the manager of all available spectrum resources. It can formulate unit bandwidth prices and trade spectrum resources with network operators in real time according to the purchase demand of network operators, so as to maximize benefits. A network operator is a provider of network services in a certain area. It can update the amount of spectrum purchased in real time according to its own user needs and unit bandwidth price, so as to maximize benefits and maximize the number of accessible devices.

The basic goal of this embodiment is to perform dynamic spectrum sharing among multiple operators through the Stackelberg game established by the spectrum provider as the leader and the network operator as the follower under the dynamic spectrum sharing mechanism, and finally realize the spectrum provider and the network. The operator's benefit is maximized, and the number of IoT devices that can be connected to the network operator is maximized.

赫兹。蜂窝用户和物联网设备适用不同的频谱分配方案，对于蜂窝用户，网络运营商以循环方式均等地调度频谱资源。对于物联网设备，每个设备从F个子信道中随机选择k个作为候选，并以p_m概率接入这k个子信道中的一个，这意味着有可能同时复用多个子信道进行传输。基于该分配方案，物联网设备选择子信道l(l∈{F})作为候选的概率为

接入该子信道的概率为p_lp_m。Some specific details of the available spectrum in the network scenario in which this embodiment is located are as follows: there is one spectrum provider, M network operators, the total available bandwidth owned by the spectrum provider is B ^tol Hz, the network operator m (m ∈ M ) purchased from a spectrum provider with a spectrum bandwidth of B _m Hz. Let the available spectrum of network operator m be divided into F sub-channels, then the bandwidth of each sub-channel is

hertz. Different spectrum allocation schemes apply to cellular users and IoT devices. For cellular users, network operators schedule spectrum resources equally in a round-robin fashion. For IoT devices, each device randomly selects k from F sub-channels as candidates, and accesses one of these k sub-channels with p _m probability, which means that it is possible to multiplex multiple sub-channels for transmission at the same time. Based on this allocation scheme, the probability that the IoT device selects the sub-channel l(l∈{F}) as a candidate is

The probability of accessing this subchannel is p _l p _m .

In order to establish the benefit function of the spectrum provider and the network operator in this embodiment, considering that the spectrum sharing in the future wireless network will be dominated by the uplink short packet communication from the terminal IoT device to the base station, it is Transport and multiplexing for modeling and analysis. In the network environment of this embodiment, cellular users and IoT devices coexist. According to the coexistence access rule: each cellular user uses orthogonal channels that do not interfere with each other, while the IoT devices randomly reuse some sub-channels for transmission; After connecting with the IoT device to the network operator, the fee will only be charged when the transmission rate exceeds a certain threshold; the network operator will give priority to ensuring the access and transmission of cellular users, and will only start to serve things after all cellular users are satisfied. Internet-connected devices. Therefore, it is necessary to first obtain the transmission rate of cellular users and IoT devices in two cases with or without IoT devices through the analysis of the following five parts, and thus obtain the cellular users and coexisting IoT devices under each network operator. The question of the maximum number of devices.

(1) When not coexisting with IoT devices, the average transmission rate of cellular users

其中P_C是蜂窝用户的发射功率，h_c是从发射机到接收机的衰落增益，d_c表示发射机到接收机的距离，η_c表示蜂窝用户的路径损耗指数，ρ_c是在信道反转功率控制模型下蜂窝用户的接收方需要接收到的信号功率阈值，W₀表示噪声功率。由此根据随机几何技术和香农公式，可得出没有物联网设备与之共存的蜂窝用户平均频谱效率

其中

是指数积分。For the case that no IoT device coexists with it, the communication transmission of cellular users will not be interfered, and the receiver signal-to-noise ratio (SNR) can be expressed as:

where P _C is the transmit power of the cellular user, h _c is the fading gain from the transmitter to the receiver, d _c is the distance from the transmitter to the receiver, η _c is the path loss exponent of the cellular user, ρ _c is the Under the transfer power control model, the receiver of the cellular user needs to receive the signal power threshold, and W ₀ represents the noise power. From this, according to the random geometry technique and Shannon's formula, the average spectral efficiency of cellular users without coexistence of IoT devices can be obtained

in

is the exponential integral.

其中N_c是网络运营商m的平均蜂窝用户数。因此，分配给一个蜂窝用户的平均频谱带宽为：

由此得出，不与物联网设备共存时，蜂窝用户的平均传输速率为：For cellular users that do not coexist with IoT devices, the network operator schedules spectrum resources in a round-robin fashion, and the probability of assigning sub-channel l to the user is

where _Nc is the average number of cellular users of network operator m. Therefore, the average spectrum bandwidth allocated to a cellular user is:

It follows from this that the average transfer rate for cellular users when not coexisting with IoT devices is:

(2) When coexisting with IoT devices, the average transmission rate of cellular users

中相同子信道l的复用，

表示复用子信道l的所有物联网设备集合。所有蜂窝用户和物联网设备的接收机都是网络运营商下的基站，则接收机受到的总干扰为

其中P_r是物联网设备X_i的发射功率，h_r表示从X_i到基站的衰落增益，d_r表示X_i到基站的距离，η_r表示物联网设备的路径损耗指数。因此接收机的信号与干扰加噪声比(SINR)可以表示为：

由此，根据香农公式可以得出与物联网设备共存的蜂窝用户的平均频谱效率

Denote by N _c and N _I as the number of cellular users and IoT devices under a network operator, respectively. When a cellular user accesses sub-channel 1, interference from other IoT devices

The multiplexing of the same subchannel l in the

Represents the set of all IoT devices that multiplex sub-channel l. The receivers of all cellular users and IoT devices are base stations under the network operator, so the total interference received by the receivers is

where _Pr is the transmit power of the _IoT device Xi, hr represents the fading gain from Xi to the base station, _dr represents the distance from _{Xi to the base station, and η r represents the} path loss index of the _IoT device. So the signal-to-interference-plus-noise ratio (SINR) of the receiver can be expressed as:

From this, the average spectral efficiency of cellular users coexisting with IoT devices can be derived according to Shannon's formula

As in the case without IoT devices, the network operator schedules spectrum resources for cellular users in the same way, so the average spectrum bandwidth allocated to each cellular user is also the above B _c . From this, the average transfer rate for cellular users when coexisting with IoT devices is:

(3) Average transmission rate of IoT devices when coexisting with cellular users

)和复用同一子信道l的其他物联网设备(表示为

)。因此，物联网设备在子信道l中遭受的总体干扰可以表示为：For IoT devices multiplexing subchannel 1, the interference comes from cellular users in subchannel 1 (denoted as

) and other IoT devices multiplexing the same sub-channel l (denoted as

). Therefore, the total interference suffered by IoT devices in sub-channel l can be expressed as:

ρ_r是在信道反转功率控制模型下物联网设备的接收方需要接收到的信号功率阈值。由此，与蜂窝用户共存的物联网设备平均频谱效率

Then the signal-to-interference-plus-noise ratio (SINR) of the receiver of the IoT device can be expressed as

ρ _r is the signal power threshold that the receiver of the IoT device needs to receive under the channel inversion power control model. From this, the average spectral efficiency of IoT devices co-located with cellular users

因此，与蜂窝用户共存时，物联网设备的平均传输速率可以表示为：According to the aforementioned resource allocation scheme, the average spectrum bandwidth allocated to an IoT device is

Therefore, the average transfer rate of an IoT device when coexisting with cellular users can be expressed as:

After obtaining the average transmission rates of cellular users and IoT devices in both cases, the relationship between the total bandwidth allocated by each network operator and the maximum coexistence of cellular users and IoT devices can be derived.

(4) Maximum number of cellular users accessing network operators when not coexisting with IoT devices

When the total bandwidth B _m purchased by the network operator m is insufficient, according to the coexistence access rule, the access of the cellular users will be guaranteed preferentially. In this case without the interference of IoT devices, let R _c,0 be the minimum transmission rate threshold of cellular users specified by the coexistence access rules, and the bandwidth B _m is known, the number of cellular users N can be obtained by solving the following problems The maximum value of _m,c :

是需要在网络运营商m中进行通信的蜂窝用户总数。通过解上述问题，可得到最大蜂窝用户接入数量函数

显然，带宽B_m越大，可接入的蜂窝用户越多。因此，对于上述问题，存在刚好能够满足所有蜂窝用户且速率刚好达到阈值R_c,0的带宽B_m,0。in

is the total number of cellular users who need to communicate in network operator m. By solving the above problems, the function of the maximum number of cellular users can be obtained

Obviously, the larger the bandwidth B _m is, the more cellular users can be accessed. Therefore, for the above problem, there is a bandwidth B _m,0 that just satisfies all cellular users and the rate just reaches the threshold R _c, 0 .

(5) Maximum number of cellular users and IoT devices coexisting

When the total bandwidth B _m purchased by the network operator m is sufficient, the IoT devices can coexist and access after all cellular users are provided with guaranteed quality services. Let R _c,0 and R _I,0 be the minimum transmission rate thresholds of cellular users and IoT devices, respectively. Knowing the total bandwidth B _m , the number of co-existing IoT devices N _m,I can be obtained by solving the following problems The maximum value of:

和

分别是需要在网络运营商m中进行通信的蜂窝用户总数和物联网设备总数。通过解上述问题，可得到物联网设备最大共存接入数量函数

in

and

are the total number of cellular users and the total number of IoT devices that need to communicate in network operator m, respectively. By solving the above problems, the function of the maximum coexistence access number of IoT devices can be obtained

After completing the analysis of the maximum number of cellular users and coexisting IoT devices supported by a single network operator under a given total bandwidth _Bm , the benefit functions _Uw and _Um of the spectrum provider and network operator can be established, respectively. Spectrum providers need to obtain spectrum use licenses from the official communications regulator, and then sublease the licensed spectrum to network operators through spectrum trading. Using a binary function to quantify the spectrum benefits obtained by the spectrum provider from the network operator, the benefit function of the spectrum provider is obtained as:

where r _p is the spectrum fee charged from all network operators, C is the license fee paid to the regulator, P is the unit bandwidth price charged by the spectrum provider, B _m is the amount of bandwidth purchased by the network operator m, and B ^tol is the total licensed spectrum bandwidth owned by the spectrum provider.

当购得的带宽B_m≥B_m0时，物联网设备可以接入网络运营商与蜂窝用户共存，且在分析(5)中已得出物联网设备的最大共存接入数量为

假设网络运营商m中蜂窝用户和物联网设备的总数分别为

和

则其效益可表示为从其客户收取的接入费用减去支付给频谱提供商的带宽费用：Each network operator charges cellular users and IoT devices access fees based on coexistence access rules. For network operator m, when the purchased bandwidth B _m ≤ B _m,0 , the bandwidth is insufficient, so only cellular users can access, and in analysis (4), it has been concluded that cellular users do not coexist with IoT devices The maximum number of access

When the purchased bandwidth B _m ≥ B _m0 , IoT devices can access network operators and coexist with cellular users, and in the analysis (5), it has been concluded that the maximum coexistence access number of IoT devices is

Suppose the total number of cellular users and IoT devices in network operator m are

and

The benefit can then be expressed as the access fee charged by its customers minus the bandwidth fee paid to the spectrum provider:

where P _c and P _I represent the access fee charged to each cellular user and IoT device, respectively.

The content of the Stackelberg game, with the spectrum provider as the leader and the network operator as the follower, includes two sub-games in the dynamic Stackelberg game running under the dynamic spectrum sharing mechanism:

(1) Spectrum provider revenue maximization sub-game: Spectrum providers continuously adjust the unit bandwidth price P, and try to provide the most suitable and most favorable price to attract network operators to purchase more spectrum bandwidth, and ultimately maximize their own benefits. The subgame of spectrum provider revenue maximization can be expressed as:

s.t.P≥0

(2) The sub-game of network operator's revenue maximization: Each network operator will continuously try to submit appropriate bandwidth purchase requirements to maximize its own benefits according to the price P announced by the spectrum provider and its own load situation. At the same time, it can be verified that since the benefits of network operators are positively related to the number of end customers connected, the maximum number of cellular users and IoT devices will be obtained when the benefits are maximized. Considering the coexistence access rule, the minimum bandwidth B _m,0 that enables all cellular users in the network operator m to access can be calculated through the analysis (4) above. The subgame is divided into two cases, which are called network operator subgame ① and network operator subgame ②. In the case of insufficient bandwidth, there is a constraint B _m ≤ B _m,0 , which is the network operator sub-game ①:

stB _m ≤B _m,0

Similarly, for the case of sufficient bandwidth, there is a constraint B _m ≥ B _m,0 , which is the network operator subgame ②:

stB _m >B _m,0

和

之后，可以得出网络运营商的最佳带宽需求

From the two network operator subgames ① and ②, respectively

and

After that, the optimal bandwidth requirements of the network operator can be derived

频谱提供商的子博弈中也存在一种最佳定价策略。最佳带宽需求策略可以通过解两个网络运营商子博弈①②得出。将最佳需求的带宽量

带入频谱提供商效益最大化子博弈，可将其重写为：Since the problem is a one-dimensional integer programming, the global optimal solution can be obtained with linear complexity by the exhaustive method. After theoretical derivation and proof, there is an optimal bandwidth demand strategy in the subgame of network operators

There is also an optimal pricing strategy in the subgame of spectrum providers. The optimal bandwidth requirement strategy can be obtained by solving the two network operator subgames ①②. The amount of bandwidth that will be optimally required

Bringing into the spectrum provider benefit maximization subgame, it can be rewritten as:

s.t.P≥0

表示在频谱价格P的条件下的最佳带宽需求策略。容易得出

相对于P单调递减。因此使用一种价格迭代的方法求解频谱提供商的子博弈：逐步提高频谱价格P，同时尽可能保证网络运营商购尽总带宽B^tol，即每次迭代中P的增量与网络运营商总需求带宽和B^tol之差成正比。价格迭代公式为：

其中t＝1,2,…是迭代次数，α_t是第t次迭代中的价格调整参数，为避免更新后的价格在最佳价格附近振荡，可以逐渐减小参数α_t来保证收敛。in

represents the optimal bandwidth demand policy under the condition of spectrum price P. easy to draw

Monotonically decreasing with respect to P. Therefore, a price iteration method is used to solve the sub-game of spectrum providers: gradually increase the spectrum price P, and at the same time ensure that the network operator purchases the total bandwidth B ^tol as much as possible, that is, the increment of P in each iteration is equal to the total bandwidth of the network operator. The required bandwidth is proportional to the difference between B ^tol . The price iteration formula is:

where t=1,2,... is the number of iterations, and α _t is the price adjustment parameter in the t-th iteration. In order to avoid the updated price from oscillating around the optimal price, the parameter α _t can be gradually reduced to ensure convergence.

During the game, if the spectrum provider adjusts the spectrum price to the best price, and the network operator calculates and submits the best bandwidth demand according to the spectrum price and load situation, then the Stackelberg game is balanced, and the spectrum provider and network operator The benefits of the network operators are maximized, and the IoT devices connected to the network operators are maximized.

The steps of the Stackelberg game, an iterative algorithm is used in the game process, and the duration of each iteration is T, then the first iteration runs in the time period [0, T], and the second iteration runs in the time period [T, 2T] ], and so on. The following steps (2) and (3) are performed at the beginning of each iteration, and after (3) is completed, the connection status of the end user, device, etc. remains unchanged for the subsequent time of this iteration. The entire game process is shown in Figure 2, and the steps are:

(1) Initialization is performed at the start of the game, that is, before the first iteration, given the stopping tolerance ε, the spectrum provider broadcasts a lower initial spectrum price P ₀ to all M network operators.

提交给频谱提供商，并通过效益函数计算出自身效益U_m。(2) Start the _t -th iteration ( _{t =} 1, 2, . . . ), each network operator solves the network operation by solving the The quotient game ① and ② get the current optimal bandwidth requirement

Submit it to the spectrum provider, and calculate its own benefit U _m through the benefit function.

更新频谱定价P_t+1。(3) The spectrum provider conducts spectrum transactions with the network operator according to the current spectrum pricing P _t and the bandwidth purchase demand submitted by the network operator, calculates its own benefit U _w through the benefit function, and uses the price iteration formula

Update spectrum pricing P _t+1 .

(4) If |P _t+1 -P _t |≤ε is established, it is considered that the Stackelberg game has reached equilibrium, the spectrum provider will not announce the new price P _t+1 , and the game is over; otherwise, the spectrum provider will pay the network operator The new price P _t+1 is announced, and the game turns to (2) to continue the t+1th iteration.

Claims (8)

1. a kind of Stackelberg game multi-operator dynamic spectrum sharing method for large-scale IoT access, is characterized in that, described multi-operator dynamic spectrum sharing method comprises the following steps: Step 1: Pre-establish the benefit function of spectrum providers and network operators; Step 2: Establish a dynamic Stackelberg game with spectrum providers as leaders and network operators as followers; Step 3: Through the game, the equilibrium is finally reached, the multi-operator spectrum sharing and the maximization of the income of the leader and the follower can be achieved, and the number of IoT devices connected to the follower can be maximized at the same time; The spectrum provider is the manager of all available spectrum resources, and can formulate unit bandwidth prices and trade spectrum resources with the network operator in real time according to the purchase demand of the network operator; the network operator is the network service provider in a small area. Providers can update spectrum purchases in real time according to their own load conditions and unit bandwidth prices; In the applicable network environment, cellular users and IoT devices exist at the same time, and the two follow the coexistence access rules when accessing the network: each cellular user uses orthogonal channels that do not interfere with each other, while IoT devices randomly reuse some sub-channels Transmission; after cellular users and IoT devices access the network operator, they will be charged only when the transmission rate exceeds a certain threshold; the network operator will give priority to ensuring the access and transmission of cellular users, and after all cellular users are satisfied will start to serve IoT devices; The multi-operator network environment is as follows: there is one spectrum provider, M network operators, the total available bandwidth owned by the spectrum provider is B ^tol Hz, and the spectrum purchased by the network operator m (m ∈ M) from the spectrum provider The bandwidth is B _m Hz; the available spectrum of network operator m is divided into F sub-channels, and the bandwidth of each sub-channel is

Hertz; different spectrum allocation schemes apply to cellular users and IoT devices. For cellular users, network operators schedule spectrum resources equally in a round-robin manner; for IoT devices, each device randomly selects k from F subchannels as candidates , and access one of the k sub-channels with p _m probability, which means that it is possible to multiplex multiple sub-channels for transmission at the same time; based on this allocation scheme, IoT devices select sub-channel l(l∈{F}) as the The candidate probability is

The probability of accessing the subchannel is p _l p _m ; The benefit function U _w of spectrum providers and network operators is expressed as:

where r _p represents the spectrum fee charged from all network operators, C is the bandwidth license fee paid to the spectrum regulator, and P is the unit bandwidth price charged by the spectrum provider; The benefit function U _m of the network operator can be expressed as:

in

and

are the maximum access number of cellular users and the maximum coexisting access number of IoT devices under the network operator m, respectively, P _c and P _I represent the access fees charged by the network operator to each cellular user and IoT device; In step 2: The dynamic Stackelberg game established with the spectrum provider as the leader and the network operator as the follower includes two sub-games: a. Spectrum providers constantly adjust the unit bandwidth price P, and try to provide the most appropriate and favorable price to attract network operators to purchase more spectrum bandwidth, and ultimately maximize their own benefits. The sub-game of spectrum provider revenue maximization is expressed as :

s.t.P≥0 b. According to the price P announced by the spectrum provider and its own load situation, the network operator continuously submits appropriate bandwidth purchase requirements to maximize its own benefits, and since the benefits are positively related to the number of connected end customers, when the benefits are maximized Obtain the maximum number of cellular users and IoT device access; according to whether the bandwidth is sufficient to meet the needs of all cellular users, the subgame is divided into two cases: ①B _m ≤B _m,0 , insufficient bandwidth:

stB _m ≤B _m,0 ②B _m ≥B _m,0 , the bandwidth is sufficient:

2. a kind of Stackelberg game multi-operator dynamic spectrum sharing method for large-scale IoT access as claimed in claim 1, is characterized in that: the Stackelberg game in described step 2 needs to carry out following steps: Step (2.1): Initialize at the beginning of the game, the spectrum provider presets a stopping tolerance, and broadcasts the preset initial spectrum pricing to all network operators; Step (2.2): Each network operator tries to submit appropriate bandwidth purchase requirements according to the spectrum price announced by the spectrum provider and the load of its own cellular users and IoT devices, so as to maximize its own benefits; Step (2.3): The spectrum provider trades spectrum with the network operator according to the bandwidth purchase requirements submitted by the network operator, and calculates and updates the spectrum pricing according to a predetermined price iteration formula; Step (2.4): If the game does not converge, the spectrum provider announces the new pricing to all network operators, and repeats steps (2.2), (2.3) and (2.4) until the game reaches a specified convergence state, At this time, the benefits of spectrum providers and network operators should be maximized, and the number of IoT devices that can access the network should also be maximized. 3. a kind of Stackelberg game multi-operator dynamic spectrum sharing method for large-scale IoT access as claimed in claim 1, is characterized in that: follower in described Stackelberg game, namely the self-load situation of network operator It is derived from the number of cellular users and IoT devices within its service area requesting access. 4. a kind of Stackelberg game multi-operator dynamic spectrum sharing method oriented towards large-scale IoT access as claimed in claim 1, is characterized in that: whether there is Internet of Things equipment in the network environment to co-exist cellular users and things in two situations. The transfer rate of a networked device is expressed as follows: a. When not coexisting with IoT devices, the average transmission rate of cellular users is

in

is the average spectral efficiency of cellular users without IoT devices coexisting with them, W ₀ represents the noise power, ρ _c is the signal power threshold that the receiver of the cellular user needs to receive under the channel inversion power control model,

is the exponential integral;

is the average spectrum bandwidth allocated to a cellular user by the network operator, and N _c represents the average number of cellular users of the network operator; b. When coexisting with IoT devices, the average transmission rate of cellular users is

in

is the average spectral efficiency of cellular users coexisting with IoT devices, N _I represents the average number of IoT devices of network operators, ρ _r represents the threshold of signal power that the receiver of IoT devices needs to receive; c. When coexisting with cellular users, the average transfer rate of IoT devices is

in

Average spectral efficiency for IoT devices co-located with cellular users,

The average spectrum bandwidth allocated to an IoT device by a network operator. 5. a kind of Stackelberg game multi-operator dynamic spectrum sharing method for large-scale IoT access as claimed in claim 1 is characterized in that: the maximum coexistence number of cellular users and IoT devices that each network operator can accommodate is the same as the The total bandwidth allocated is related as follows: a. When not coexisting with IoT devices, the maximum number N _m,c of cellular users accessing network operators can be obtained by solving the following problems:

in

is the total number of cellular users who need to communicate in the network operator m; N _m,c and the bandwidth B _m are in a mono-increasing relationship, there is a bandwidth B _m that can just satisfy all cellular users and the rate just reaches the charging threshold rate R _{c,0 ,0} ; b. When coexisting with cellular users, the maximum number N _m,I of IoT devices accessing the network operator is obtained by solving the following problems:

in

is the total number of IoT devices that need to communicate in network operator m. 6. a kind of stackelberg game multi-operator dynamic spectrum sharing method for large-scale IoT access as claimed in claim 2, is characterized in that: the suitable optimal bandwidth purchase requirement that network operator m submits in game process

By solving the network operator subgame ①②, we get:

in

and

are the maximum benefits in the two cases solved by the network operator sub-games ① and ② respectively. 7. a kind of Stackelberg game multi-operator dynamic spectrum sharing method for large-scale IoT access as claimed in claim 2, it is characterized in that: spectrum operator sub-game is solved by a kind of method of price iteration, namely gradually increasing spectrum The price P also ensures that the network operator purchases the total bandwidth B ^tol as much as possible. The increment of P in each iteration is proportional to the difference between the network operator's total demanded bandwidth and B ^tol . The price iteration formula is:

where t=1,2,... is the number of iterations,

Represents the optimal bandwidth purchase demand of network operators under the condition of spectrum price P _t , α _t is the price adjustment parameter in the t-th iteration, α _t gradually decreases during the iteration process to avoid the updated price at the optimal price oscillate nearby. 8. a kind of stackelberg game multi-operator dynamic spectrum sharing method for large-scale IoT access as claimed in claim 2, is characterized in that: the condition that judges Stackelberg game to reach equilibrium is |P _t+1 -P _t |≤ ε is established, and ε is the pre-set stopping tolerance before the game starts. If this condition is satisfied, the game is considered to reach an equilibrium convergence state and the game ends.

Images