A resource allocation method based on game theory and genetic algorithm

Abstract

The present invention discloses a resource allocation method based on game theory and genetic algorithm, including: constructing a communication model of communication equipment and obtaining the channel capacity from the base station to the user communication equipment; constructing a game model of the optimal channel access strategy according to the channel capacity from the base station to the user communication equipment; obtaining the optimal solution of the game model based on the genetic algorithm as the result of resource allocation. The present invention models interference and user communication equipment as both parties in the game, so that it can dynamically respond to complex and changeable electromagnetic environments. Compared with the shortcomings of traditional algorithms that perform poorly under high load and multi-device conditions, the resource allocation method of the present invention can better adapt to complex scenarios, thereby improving the efficiency and reliability of spectrum management.

Description

Technical Field

[0001] The invention belongs to the technical field of resource allocation, and in particular relates to a resource allocation method based on game theory and genetic algorithm. Background Art

[0002] Spectrum resource management methods refer to the methods of planning, coordinating and controlling the electromagnetic spectrum used by combining administrative and technical means with sensing equipment, suppression equipment, operations and event handling personnel. This can effectively avoid electromagnetic interference between users and limit and handle malicious interference devices to the maximum extent. Spectrum resource management was initially only about planning frequency-using equipment to find the frequency-using scheme with the least self-interference, which played a major role at the time when malicious interference technology was not well developed. With the increase in frequency-using equipment and the impact of malicious interference, a single frequency-using plan cannot adapt to the dynamically changing spectrum situation, and the increase in users has further increased the complexity of the allocation plan. Therefore, some scholars have proposed the cognitive radio (CR) architecture, which adds perception capabilities to spectrum resource management. The main idea is to use spectrum perception as the basis for spectrum resource management. It does not require additional spectrum bandwidth, but only makes reasonable use of the authorized but underutilized frequency bands. That is, when the primary users (PU) do not use the authorized frequency band, the secondary users (SU) can temporarily use the frequency band, and when the PU wants to use the authorized frequency band, the SU can stop using the frequency band in time without affecting the PU. This greatly increases the frequency band utilization, but the selection of a suitable spectrum perception method is the key to realizing the entire spectrum allocation process. Energy detection has long been the most commonly used method in this field because it does not require any prior feature information about user signals. Some scholars use the discrete-time Markov decision process and the equivalence class transformation (Eclat) algorithm to improve data transmission security and thus realize energy detection. Some scholars first perform spectrum perception through energy detection, and then optimize power allocation in combination with relay technology.

[0003] However, the above management method does not take into account the situation of external interference, so that the above system is still helpless in the face of malicious interference. Therefore, some scholars began to further explore the information in the perception data, obtain information such as channel utilization, signal-to-interference-to-noise ratio, and combine it with the transmission task requirements to better manage spectrum resources. For example, some scholars have improved the threshold of the average energy perception algorithm and added an adaptive threshold, which reduces the probability of decision errors, reduces the amount of calculation, and increases the performance of energy detection technology under low signal-to-interference-to-noise ratio conditions; some scholars have used a variety of genetic algorithms to solve multi-objective optimization problems such as spectrum resource management, and have evaluated performance from aspects such as channel utilization, signal-to-noise-to-interference ratio, and channel access delay. However, the above algorithm has the problem of high computational complexity, and there are intelligent interference devices in the environment, such as forwarding interference, smart interference, and even interference using intelligent interference strategies. The key to the above problems lies in how to efficiently allocate users' spectrum resources while monitoring and finding suspicious malicious interference devices in real time.

[0004] With the vigorous development of artificial intelligence technology, by increasing the number of neural network layers, neural network technology has shown superiority in preventing deceptive interference and finding potential interference relationships, and the ability of spectrum resource management in anti-interference and spectrum utilization has been continuously improved. Some scholars have mined and utilized the medium- and long-term spectrum characteristics in spectrum time series data based on deep learning long- and short-term memory models to improve the performance of spectrum prediction. Some scholars have taken into account the distributed characteristics of blockchain and applied blockchain technology to the spectrum sharing of large-scale ultra-dense mobile Internet, connecting massive personal wireless devices to form a spectrum device network, defining "spectrum coins" as rewards for devices to collect spectrum data, and proposed a distributed consensus mechanism consisting of perception node consensus fusion, verification node consensus verification, and cluster head node consensus confirmation.

[0005] Therefore, spectrum resource management has changed from traditional manual planning to dynamic planning; from static exclusive use to dynamic sharing. However, the above technical solutions still face many problems.

[0006] First, the current simple information perception method leads to a single spectrum space modeling representation method, which is difficult to adapt to complex electromagnetic environments. The most commonly used spectrum space modeling representation method is the spectrum waterfall diagram, which mainly focuses on the space-time-frequency resource distribution of the spectrum space. When the number of devices in the environment is small, it is relatively intuitive. With the increase in the number of controlled devices, the electromagnetic spectrum space is increasingly complex, making the spectrum waterfall diagram unable to represent complex electromagnetic systems composed of multiple entities and multiple variables.

[0007] Secondly, current spectrum management relies heavily on the operator’s experience, and the system has a low degree of automation, which cannot assist operators in making decisions and management. When the number of devices is small and there are not many limiting factors, operators can still complete management. However, once it involves managing a large number of devices or multiple users encountering problems that require decision-making and management, the huge amount of computing and judgment required will greatly reduce the robustness and real-time nature of management.

[0008] Finally, the current spectrum management methods between different devices are different, and information between devices cannot be communicated. When interference occurs, it is impossible to comprehensively consider the communication conditions of different types of equipment. We can only rely on the experience of operators and reports from equipment users, which is fatal and dangerous in major situations. Summary of the invention

[0009] In order to solve the above problems existing in the prior art, the present invention provides a resource allocation method based on game theory and genetic algorithm. The technical problem to be solved by the present invention is achieved through the following technical solutions:

[0010] The present invention provides a resource allocation method based on game theory and genetic algorithm, comprising:

[0011] S1: Build a communication model of the communication device and obtain the channel capacity from the base station to the user communication device;

[0012] S2: constructing a game model of an optimal channel access strategy according to the channel capacity from the base station to the user communication device, wherein the game model includes a sub-game model based on a leader level of interference devices and a sub-game model based on a follower level of user communication devices;

[0013] S3: Obtaining the optimal solution of the game model based on the genetic algorithm as the result of resource allocation.

[0014] In one embodiment of the present invention, the S1 includes:

[0015] S1.1: Construct the signal path loss and signal fading model of user communication equipment;

[0016] S1.2: Obtain an expression for the signal received by the user communication device at the base station and an expression for the channel capacity. In one embodiment of the present invention, the expression for the signal path loss is:

[0017] l BS,n =32.45+20lgd BS,n +20lgf-G t -G r

[0018] Among them, l BS,n represents the signal path loss between the base station and the nth user communication device, Gt represents the base station transmitting antenna gain, G r represents the receiving antenna gain of the user communication device, d BS,n represents the straight-line distance between the base station and the nth user communication device, f represents the frequency of the base station transmitting signals, 1≤n≤N, and N represents the total number of user communication devices;

[0019] The signal fading model of the user communication device is expressed as:

[0020]

[0021] Among them, h rician represents the Rice fading channel, β is the Rice factor, which represents the strength ratio of the direct link to the equivalent multiple reflection links; h LoS represents a direct link, h NLoS Indicates an indirect link.

[0022] In one embodiment of the present invention, the S1.2 includes:

[0023] Get the expression of the signal received by the user communication device from the base station:

[0024]

[0025] Among them, y n Indicates that the nth user communication device receives the signal from the base station, l BS,n represents the signal path loss between the base station and the nth user communication device, P n represents the signal power from the base station to the nth user communication device; x BS Indicates the signal sent by the base station, A n represents the communication frequency of the nth user communication device, A n′ A represents the communication frequency of the n′th user communication device; j represents the interference strategy of the jth interference device, that is, the interference frequency band of the jth interference device, 1≤j≤J, J represents the number of interference devices, δ(A n ,A n′ ) represents the indicator function, A n =A n′ Then δ(A n ,A n′ ) takes the value of 1, A n ≠A n′ Then δ(A n ,A n′ ) takes the value of 0; δ(A n ,A j ) represents the indicator function, A n =A j Then δ(A n ,Aj ) takes the value of 1, A n ≠A j Then δ(A n ,A j ) takes the value of 0; n′ is the signal power from the base station to the n′th user communication device, l n′,n represents the signal path loss between the nth user communication device and the n′th user communication device; P j represents the signal power of the interfering device, l j,n represents the signal path loss between the interference device j and the nth user communication device, x j represents the signal emitted by the interference device, η represents the mean value is 0, and the variance is σ 2 Additive Gaussian white noise;

[0026] The expression for the channel capacity from the base station to the user communication device is obtained:

[0027]

[0028] in, represents the channel capacity from the base station to the nth user communication device in the current resource allocation scheme, and B represents the communication bandwidth.

[0029] In one embodiment of the present invention, the S2 includes:

[0030] S2.1: According to the Stackelberg game model, let the interference device be the leader and the user communication device be the follower;

[0031] S2.2: Construct a follower-level sub-game model to find a channel access strategy with the highest base station to device throughput The sub-game model at the follower level is expressed as:

[0032] g U = {N,c,C n (A n ,A n′ ,A J )},n,n′∈N,n′≠n,j∈J,

[0033] Where c represents the channel;

[0034] S2.3: Constructing a sub-game model at the leader level to pursue a jamming strategy that maximizes the jamming utility of the jamming device The sub-game model at the leader level is expressed as:

[0035] g JAM ={J,c,U J (A n ,Aj )},n∈N,j∈J,

[0036] in, Indicates the jamming effectiveness of the jamming device;

[0037] S2.4: A multi-layer Stackelberg game model is constructed based on the sub-game model of the follower level and the sub-game model of the leader level, which is expressed as:

[0038]

[0039] In one embodiment of the present invention, S3 includes:

[0040] S3.1: Two initial genetic populations are constructed by random initialization, and the individual fitness functions of the two initial genetic populations are defined as:

[0041]

[0042] S3.2: The genotype of individuals in the population is defined by integer coding. The genotype of each individual includes the frequency and discrete power of the channel resource.

[0043] S3.3: Evaluate the fitness value of each individual in the population, and perform selection, crossover and mutation operations on the individuals in the population according to the fitness value;

[0044] S3.4: Obtain the fitness of each individual in the population after selection, crossover and mutation operations and determine whether the number of iterations has been reached. If not, re-execute step S3.3. If so, select the individual with the largest fitness in the current population as the optimal solution for resource allocation.

[0045] In one embodiment of the present invention, in step S3.3, a proportional selection and optimal individual retention method is used to select individuals in the population.

[0046] In one embodiment of the present invention, in step S3.3, an adaptive crossover probability is used to perform a crossover operation on individuals in the population, and the expression of the adaptive crossover probability is:

[0047]

[0048] Among them, p c represents the adaptive crossover probability, and Respectively represent the maximum and minimum values ​​of all adaptive crossover probabilities, f i represents the fitness of individual i, f max represents the maximum value of fitness, f avg represents the average value of fitness;

[0049] The adaptive mutation probability is used to perform crossover operation on individuals in the population. The expression of the adaptive mutation probability is:

[0050]

[0051] Among them, p m represents the adaptive mutation probability, and They represent the maximum and minimum values ​​of the adaptive mutation probability respectively.

[0052] Another aspect of the present invention provides a storage medium storing a computer program for executing the steps of the resource allocation method based on game theory and genetic algorithm described in any one of the above embodiments.

[0053] Another aspect of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor calls the computer program in the memory, it implements the steps of the resource allocation method based on game theory and genetic algorithm as described in any of the above embodiments.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1. Higher adaptability: The resource allocation method based on game theory and genetic algorithm proposed in the present invention models interference and legitimate user communication devices as the two sides of the game, so that it can dynamically respond to complex and changing electromagnetic environments. Compared with the shortcomings of traditional algorithms that perform poorly under high load and multi-device conditions, the resource allocation method of the present invention can better adapt to complex scenarios, thereby improving the efficiency and reliability of spectrum management.

[0056] 2. Optimize resource allocation: User communication equipment uses genetic algorithms to find the optimal resource allocation solution. This approach ensures that resources in the frequency domain and power domain can be used more reasonably and efficiently, overcomes the shortcomings of existing algorithms in unreasonable allocation in complex environments, and helps achieve better spectrum resource utilization.

[0057] 3. Enhanced robustness: The resource allocation method based on game theory and genetic algorithm proposed in the present invention improves the robustness of the system in dealing with complex electromagnetic systems composed of multiple variables and multiple entities through the framework of game theory and the optimization ability of genetic algorithm. Compared with the existing algorithms that are prone to reduce management robustness under high computing requirements, the method of the present invention can provide more real-time resource allocation decisions while maintaining system stability.

[0058] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is a flow chart of a resource allocation method based on game theory and genetic algorithm provided by an embodiment of the present invention;

[0060] Figure 2 It is a schematic diagram of the genotype of a population individual provided by an embodiment of the present invention;

[0061] Figure 3 is a genetic algorithm flow chart provided by an embodiment of the present invention;

[0062] Figure 4 is a schematic diagram of a simulation scenario provided by an embodiment of the present invention;

[0063] Figure 5 It is a schematic diagram of the iterative process of the channel selection probability of the interference device in a certain round of training;

[0064] Figure 6 It is a schematic diagram of the iterative process of the channel selection probability of the user communication device in a certain round of training;

[0065] Figure 7 It is a curve chart showing the change of interference effectiveness of each interference device during the whole game process;

[0066] Figure 8 It is a comparison chart of average channel performance under different resource allocation methods. DETAILED DESCRIPTION

[0067] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, a resource allocation method based on game theory and genetic algorithm proposed in accordance with the present invention is described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0068] The above and other technical contents, features and effects of the present invention are clearly presented in the following detailed description of the specific implementation modes in conjunction with the accompanying drawings. Through the description of the specific implementation modes, the technical means and effects adopted by the present invention to achieve the predetermined purpose can be more deeply and specifically understood. However, the attached drawings are only for reference and explanation purposes and are not used to limit the technical solutions of the present invention.

[0069] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants are intended to cover non-exclusive inclusion, so that an article or device including a series of elements includes not only those elements, but also other elements that are not explicitly listed. In the absence of more restrictions, the elements defined by the statement "including one..." do not exclude the existence of other identical elements in the article or device including the elements.

[0070] Embodiment 1

[0071] See also Figure 1 , Figure 1 1 is a flow chart of a resource allocation method based on game theory and genetic algorithm provided by an embodiment of the present invention. The resource allocation method specifically includes the following steps:

[0072] S1: Construct a communication model of the user communication device and obtain the channel capacity from the base station to the user communication device.

[0073] Step S1 of this embodiment specifically includes:

[0074] S1.1: Construct the signal path loss and signal fading model of the user communication equipment.

[0075] First, the communication model of the user communication equipment in the field is defined. The communication model is divided into a signal path loss and a signal fading model. The signal path loss of the user communication equipment in the open space is expressed as:

[0076] l BS,n =32.45+20lgd BS,n +20lgf-G t -G r (dB) (1)

[0077] Among them, l BS,n represents the signal path loss between the base station and the nth user communication device, G t represents the base station transmitting antenna gain, G r represents the receiving antenna gain of the user communication device, d BS,n represents the straight-line distance between the base station and the nth user communication device (Km), f represents the frequency of the base station transmitting signal (Mhz), 1≤n≤N, and N represents the total number of user communication devices.

[0078] Assuming that the communication channel of the user communication device is a LoS path, a Rice fading channel is adopted, and the signal fading model of the user communication device is expressed as:

[0079]

[0080] Among them, h rician represents the Rice fading channel, β is the Rice factor, which represents the strength ratio of the direct link to the equivalent multiple reflection links. When there is no direct link, the Rice fading will degenerate into Rayleigh fading; h LoS represents a direct link, which is related to the arrival angle of the signal. The mathematical model of this embodiment adopts a single antenna, so h here LoS =1,h NLoS Indicates an indirect link.

[0081] S1.2: Obtain the expression of the signal received by the user communication device from the base station and the expression of the channel capacity.

[0082] Specifically, the expression of the signal received by the user communication device from the base station is obtained:

[0083]

[0084] Among them, y n Indicates that the nth user communication device receives the signal from the base station, l BS,n represents the signal path loss between the base station and the nth user communication device, P n represents the signal power from the base station to the nth user communication device; x BS Indicates the signal sent by the base station, A n represents the communication frequency of the nth user communication device, A n′ A represents the communication frequency of the n′th user communication device; j represents the interference strategy of the jth interference device, that is, the interference frequency band of the jth interference device, 1≤j≤J, J represents the number of interference devices, δ(A n ,A n′ ) represents the indicator function, A n =A n′ Then δ(A n ,A n′ ) takes the value of 1, A n ≠A n′ Then δ(A n ,A n′ ) takes the value of 0; δ(A n ,A j ) represents the indicator function, A n =A j Then δ(A n ,A j ) takes the value of 1, A n ≠A j Then δ(A n ,A j ) takes the value of 0;n′ is the signal power from the base station to the n′th user communication device, l n′,n represents the signal path loss between the nth user communication device and the n′th user communication device; P j represents the signal power of the interfering device, l j,n represents the signal path loss between the interference device j and the nth user communication device, x j represents the signal emitted by the interference device, η represents the mean value is 0, and the variance is σ 2 Additive white Gaussian noise.

[0085] Therefore, the expression of the channel capacity from the base station to the user can be further obtained as follows:

[0086]

[0087] Among them, C n (A n ,A n′ ,A j ) represents the channel capacity from the base station to the nth user communication device in the current resource allocation scheme, and B represents the communication bandwidth.

[0088] S2: Constructing a game model of an optimal channel access strategy according to the channel capacity from the base station to the user communication device.

[0089] Step S2 of this embodiment specifically includes:

[0090] S2.1: According to the Stackelberg game model, let the interference device be the leader and the user communication device be the follower.

[0091] This embodiment adopts the Stackelberg game model to find the optimal channel access strategy under interference. According to the Stackelberg game model, the interference device is the leader and the user communication device is the follower. The follower first responds to the leader's strategy, and the leader will further make a new strategy based on the current one, and finally the system reaches a balance through iteration.

[0092] S2.2: Construct a sub-game model at the follower level.

[0093] This part of the game is to find a channel access strategy that minimizes malicious interference, that is, to find a channel access strategy that maximizes the throughput from the base station to the user communication device. The channel access strategy is expressed as:

[0094]

[0095] Therefore, the sub-game model at the follower level can be expressed as:

[0096] g U = {N,c,C n (A n ,A n′ ,A J )},n,n′∈N,n′≠n,j∈J (6)

[0097] Here, c represents the channel.

[0098] S2.3: Construct a subgame model at the leader level.

[0099] For the interference device set J, its interference strategy A J It can be expressed as: pursuing the maximization of the impact caused by interference, so the interference utility U of the interference device is defined as J and the jamming strategies of the jamming devices

[0100]

[0101] Similarly, the sub-game model at the leader level can be expressed as:

[0102] g JAM ={J,c,U J (A n ,A j )},n∈N,j∈J (9)

[0103] In summary, the multi-layer Stackelberg game model can be expressed as:

[0104]

[0105] The Stackelberg game has a unique optimal solution. Therefore, the communication-interference system can achieve the optimal solution of the Stackelberg game through an iterative algorithm, that is, neither the leader nor the follower can obtain the better effect they want by changing their own strategies. That is, the expression of the optimal solution of the multi-layer Stackelberg game model can be expressed as follows:

[0106]

[0107] S3: Obtaining the optimal solution of the game model based on the genetic algorithm as the result of resource allocation.

[0108] According to the goal of frequency allocation and the fact that traditional optimization problems are prone to falling into local optimal solutions, this embodiment adopts a genetic algorithm and improves it to meet the requirements of the frequency scheduling scenario. Genetic algorithm is a heuristic optimization algorithm that simulates Darwin's theory of evolution and the "survival of the fittest" mechanism in nature to search for the global optimal solution. The genetic algorithm starts with a population of the problem (a part of the feasible solution set). Each individual in the population corresponds to a chromosome in biological evolution, and its encoding corresponds to the individual characteristics, namely the genotype. In the genetic algorithm, the decoded individual genotype is evaluated by the fitness function. In each evolutionary iteration, selection, crossover and mutation operations are used to perform parallel searches on multiple individuals in the population according to the fitness results to generate a new population. Through continuous iteration of the above process, after reaching the specified number of iterations, decoding the individual with the highest fitness is the optimal solution to the problem.

[0109] Coding is the first problem to be solved when using genetic algorithms to solve problems, and it is also a key step. Its purpose is to map the individual phenotype to the individual genotype. The coding method affects the operation method of the genetic operator and also determines the efficiency of genetic evolution to a large extent. Coding methods can be divided into three categories: binary coding, floating point coding, and symbolic coding. The role of the fitness function is to evaluate the quality of the individual solution through individual characteristics. The fitness function is generally closely related to the objective function constructed in the problem model. Since the roulette-based selection operator needs to map the fitness function to probability, it is generally required to be positive.

[0110] When evaluating the fitness of each individual, the genotype must first be decoded to obtain the phenotype, and then the objective function can be used to calculate the evaluation value of the phenotype, and finally the evaluation value can be converted into a fitness value.

[0111] Taking into account the characteristics of genetic algorithms such as easy premature maturity and relatively high time complexity, researchers have proposed many ways to improve them, one of which is based on adaptive crossover and mutation probability. Therefore, the embodiment of the present invention introduces one-dimensional local search based on adaptive crossover and mutation probability, thereby enhancing the ability of the algorithm to escape from the local optimal solution.

[0112] Specifically, step S3 of this embodiment includes:

[0113] S3.1: Use random initialization to construct two initial genetic populations, and define the individual fitness functions of these two initial genetic populations as follows:

[0114]

[0115] S3.2: The genotypes of individuals in the population are defined by integer coding. The genotype of each individual includes channel resources and discrete power.

[0116] This step defines the genotype of individuals in the population, and the encoding method used is integer encoding. The decoding method is the inverse process of encoding. Take the case where the number of channel resources is M and the discrete power selection is E as an example. For the communication function, the number of optional resource combinations is M*E, see Figure 2 , Figure 2 It is a schematic diagram of the genotypes of individuals in a population provided in an embodiment of the present invention.

[0117] S3.3: Evaluate the fitness value of each individual in the population, and perform selection, crossover and mutation operations on the individuals in the population according to the fitness value.

[0118] In genetic algorithms, the selection operation is to select offspring individuals from the parent population with a certain strategy, retaining the individuals with the best fitness while also considering population diversity to avoid falling into the local optimal solution. The individuals selected by the selection operator are further used for subsequent crossover and mutation operations. Commonly used methods include roulette wheel selection, stochastic tournament selection, and other methods.

[0119] The crossover operation of the genetic algorithm refers to exchanging genes at some positions of two paired parent individual chromosomes in a certain way. Mutation is to replace the gene values ​​at certain positions in the coding string of the individual chromosome with genes in the gene library. The crossover operator has become the main operator in the genetic algorithm due to its global search capability. The mutation operator maintains the diversity of the population through local mutation. The genetic algorithm improves the balanced search capability of the algorithm through crossover and mutation operations. The main methods for implementing the crossover operator include single-point crossover and uniform crossover, and the main methods for implementing the mutation operation include basic bit mutation and uniform mutation.

[0120] Genetic algorithm is a widely used heuristic algorithm. There are many research results in resource allocation algorithm. Through the design of encoding method, it can be used to solve the combinatorial optimization problem of discrete resources. The flowchart of using genetic algorithm to solve is as follows Figure 3 shown.

[0121] Considering that genetic algorithms are prone to premature maturity and have relatively high time complexity, researchers have proposed many ways to improve them, one of which is based on adaptive crossover and mutation probability. Therefore, the embodiment of the present invention introduces one-dimensional local search based on adaptive crossover and mutation probability to enhance the ability of the algorithm to jump out of the local optimal solution. The following describes the selection of selection operators, adaptive changes of crossover and mutation operators, and one-dimensional local search.

[0122] Specifically, the selection operation is the first step of the genetic algorithm evolution, which selects excellent individuals from the parent generation with a certain strategy, with the purpose of providing excellent genes for the next generation. In the embodiment of the present invention, the scheme of proportional selection and optimal individual retention is used to select individuals from the population. First, the individuals in the population are arranged in order of fitness, and the fitness f is used to select the best individuals. i Determine the probability of an individual being selected. The probability of individual i being selected is calculated as follows:

[0123]

[0124] The cumulative probability of each individual is calculated as follows:

[0125]

[0126] Using random numbers, if the random number at the current individual falls within the cumulative probability range of the previous individual and the current individual, then the individual is selected, and the number of selections is the same as the population size.

[0127] Furthermore, the crossover operation is an important step in the genetic algorithm, and the crossover population is generated by mixing and inheriting the characteristics of the original population. Through the crossover operation, the search range in the feasible solution space is continuously expanded, the diversity of the population is increased, and excellent individuals with higher fitness are continuously generated, thereby improving the optimization performance of the algorithm. The size of the crossover probability will affect the search ability of the genetic algorithm. If the value is too large, it may evolve into a random search, and the overall efficiency will be affected. If the value is very small, it may cause the algorithm to be slow and inefficient during the optimization process, and the global optimization ability of the algorithm is greatly reduced. In order to overcome the above problems, the present invention adopts an adaptive crossover probability to dynamically adjust the individual fitness value, the optimal fitness of the group and the average fitness in the group. When the individual fitness value is greater than or equal to the average fitness, the crossover probability is reduced and the local search ability is enhanced. Otherwise, the maximum crossover probability is used to determine whether the individual crosses, thereby enhancing the global optimization ability of the algorithm. The adjusted adaptive crossover probability is shown in the following formula.

[0128]

[0129] Among them, p c represents the adaptive crossover probability, and Respectively represent the maximum and minimum values ​​of all adaptive crossover probabilities, f i represents the fitness of individual i, f max represents the maximum value of fitness, f avg Represents the average value of fitness.

[0130] The algorithm uses two-point crossover and adaptively determines the crossover probability based on the fitness value. If the generated random number is less than the crossover probability, a crossover operation is performed.

[0131] Mutation operation is an important component of genetic algorithm and plays an indispensable role in the evolutionary process. Mutation operation is usually used in conjunction with crossover operation to enhance the global optimization ability of genetic algorithm. In fact, the performance of genetic algorithm will eventually be more sensitive to mutation probability. Appropriate mutation probability will bring extensive genetic diversity to the population, but excessive mutation probability will have a negative impact on the original individuals, which makes the genetic algorithm a complete random search process and the performance advantage of the genetic algorithm will disappear. The mutation probability can characterize the intensity of the mutation. In order to correctly force the individuals in the current population to jump out of the local optimal solution, the mutation probability needs to be dynamically adjusted. Therefore, the present invention introduces an adaptive mutation probability. When the individual fitness value is greater than or equal to the average fitness, the mutation probability is reduced and the local search ability is enhanced. Otherwise, the maximum mutation probability is used to determine whether the individual crosses over to enhance the global search ability. The adjusted adaptive mutation probability is expressed as follows:

[0132]

[0133] Among them, p m represents the adaptive mutation probability, and They represent the maximum and minimum values ​​of the adaptive mutation probability respectively.

[0134] The algorithm uses multi-point mutation, which is dynamically adjusted according to the encoding length. First, the gene number to be mutated in the individual is generated, and the mutation probability is adaptively determined according to the fitness value. If the generated random number is less than the mutation probability, the chromosome is subjected to multi-point random mutation.

[0135] S3.4: Obtain the fitness of each individual in the population after selection, crossover and mutation operations and determine whether the number of iterations has been reached. If not, re-execute step S3.3. If so, select the individual with the largest fitness in the current population as the optimal solution for resource allocation.

[0136] The effect of the resource allocation method based on game theory and genetic algorithm of the present invention is further described below through simulation experiments.

[0137] First, a simulation scenario was constructed. The entire scenario range was set within 4000m*4000m. The upper right corner was the location of the base station. Three interference devices were placed in the remaining three corners around the scenario. The user communication devices were randomly selected from multiple locations, such as Figure 4As shown, Jammer represents the jammer device, Jammer Link represents the jammer device link, Comm.Devices represents the user communication device, and BS Link represents the user communication device link. The number of available channels is 4; in the initial state, the probability of the user communication device and the jammer device selecting the channel is equal.

[0138] See also Figure 5 and Figure 6 , Figure 5 It is a schematic diagram of the iterative process of the channel selection probability of the interference device 1 in a certain round of training; Figure 6 This is a schematic diagram of the iterative process of the channel selection probability of user communication device 1 in a certain round of training, where channel represents the channel. It can be seen that after 200 rounds of iterations, the channel probability selected by each device has approached 1, so the interference device 1 will eventually choose channel 4 for interference; user 1 will eventually choose channel 1 for communication.

[0139] See also Figure 7 , Figure 7 It is a curve chart showing the change of interference efficiency of each interference device during the whole game process. It can be seen that after a period of game, the efficiency of the interference device no longer improves. At this time, it is the optimal channel selection strategy of the interference device, which is also the optimal channel selection strategy of the user communication device under interference. Therefore, it can be considered that the game has reached equilibrium.

[0140] See also Figure 8 , Figure 8 This is a comparison chart of average channel performance under different resource allocation methods, showing the change of average channel performance of users with different anti-interference methods as the number of users changes when there are 3 interfering devices. The horizontal axis represents the number of user communication devices, that is, the number of users, and the vertical axis represents the average channel performance. It can be seen that in the absence of malicious interference (No Jammer), the channel performance first maintains and then gradually decreases as the number of users increases. This is because when the number of users is less than or equal to 4, the channel is sufficient and there will be no co-channel interference, so the best channel performance is achieved; however, as the number of users increases, the co-channel interference between different users becomes more influential, causing the average channel performance to decrease.

[0141] When the proposed method is used, the channel performance increases first and then decreases. This is because when the number of users is 2, 3 interference devices can interfere with all user channels; as the number of users increases, the interference devices need to iterate to find the optimal interference strategy, so there will always be users who are not maliciously interfered by the interference devices; but as the number of users increases, malicious interference and co-channel interference jointly affect the average channel performance, resulting in a continuous decrease in performance; at the same time, it can be observed that because the random channel selection algorithm (Random method) does not select the optimal strategy, its average channel performance is not as good as the proposed method in all cases.

[0142] The present invention studies the channel resource allocation problem, takes into account the mutual interference between users and the external interference of the interfering equipment, establishes a multi-layer Stackelberg game model for the proposed communication model, and uses a genetic algorithm to solve the model to obtain the best interference strategy and communication channel selection strategy. The anti-interference performance of the method of the present invention and the existing random selection algorithm are compared. The simulation results show that the method proposed by the present invention has better performance, can adapt to more complex and harsh environments, and has more comprehensive anti-interference performance.

[0143] In summary, the resource allocation method based on game theory and genetic algorithm proposed in the present invention models noise interference and legitimate user communication equipment as the two sides of the game, so that it can dynamically respond to complex and changeable electromagnetic environments. Compared with the shortcomings of traditional algorithms that perform poorly under high load and multi-device conditions, the resource allocation method of the new algorithm of the present invention can better adapt to complex scenarios, thereby improving the efficiency and reliability of spectrum management. User communication equipment and legitimate equipment use genetic algorithms to find the optimal resource allocation plan. This approach ensures that resources in the frequency domain and power domain can be used more reasonably and efficiently, overcomes the shortcomings of existing algorithms in unreasonable allocation in complex environments, and helps to achieve better spectrum resource utilization.

[0144] In addition, the new resource allocation method based on game theory and genetic algorithm proposed in the present invention improves the robustness of the system in dealing with complex electromagnetic systems composed of multiple variables and multiple entities through the framework of game theory and the optimization ability of genetic algorithm. Compared with the existing algorithm, which is prone to reduce management robustness under high computing requirements, this new algorithm method of the present invention can provide more real-time resource allocation decisions while maintaining system stability.

[0145] Another embodiment of the present invention provides a storage medium, wherein a computer program is stored in the storage medium, and the computer program is used to execute the steps of the resource allocation method based on game theory and genetic algorithm described in the above embodiment. Another aspect of the present invention provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor calls the computer program in the memory, the steps of the resource allocation method based on game theory and genetic algorithm described in the above embodiment are implemented. Specifically, the above-mentioned integrated module implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium, including several instructions for enabling an electronic device (which can be a personal computer, a server, or a network device, etc.) or a processor (processor) to execute some steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0146] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.

Claims

1. A resource allocation method based on game theory and genetic algorithm, characterized in that: include: S1: Build a communication model of the communication device and obtain the channel capacity from the base station to the user communication device; S2: constructing a game model of an optimal channel access strategy according to the channel capacity from the base station to the user communication device, wherein the game model includes a sub-game model based on a leader level of interference devices and a sub-game model based on a follower level of user communication devices; S3: Obtaining the optimal solution of the game model based on the genetic algorithm as the result of resource allocation, The S1 includes: S1.1: Construct the signal path loss and signal fading model of user communication equipment; S1.2: Obtain the expression of the signal received by the user communication device from the base station and the expression of the channel capacity; The S2 includes: S2.1: According to the Stackelberg game model, let the interference device be the leader and the user communication device be the follower; S2.2: Construct a follower-level sub-game model to find a channel access strategy with the highest base station to device throughput The sub-game model at the follower level is expressed as: g U ={N,c,C n (A n ,A n′ ,A J )},n,n′∈N,n′≠n,j∈J, Where c represents the channel; S2.3: Constructing a sub-game model at the leader level to pursue a jamming strategy that maximizes the jamming utility of the jamming device The sub-game model at the leader level is expressed as: g JAM ={J,c,U J (A n ,A j )},n∈N,j∈J, in, Indicates the jamming effectiveness of the jamming device; S2.4: A multi-layer Stackelberg game model is constructed based on the sub-game model of the follower level and the sub-game model of the leader level, which is expressed as: g={N,J,c,C Un (A n ,A n′ ,A J ),U J (A n ,A j )},n,n′∈N,n′≠n,j∈J; The S3 includes: S3.1: Two initial genetic populations are constructed by random initialization, and the individual fitness functions of the two initial genetic populations are defined as: S3.2: The genotype of individuals in the population is defined by integer coding. The genotype of each individual includes the frequency and discrete power of the channel resource. S3.3: Evaluate the fitness value of each individual in the population, and perform selection, crossover and mutation operations on the individuals in the population according to the fitness value; S3.4: Obtain the fitness of each individual in the population after selection, crossover and mutation operations and determine whether the number of iterations has been reached. If not, re-execute step S3.

3. If so, select the individual with the largest fitness in the current population as the optimal solution for resource allocation.

2. The resource allocation method based on game theory and genetic algorithm according to claim 1 is characterized in that: The expression of the signal path loss is: l BS,n =32.45+20lgd BS,n +20lgf-G t -G r Among them, l BS,n represents the signal path loss between the base station and the nth user communication device, G t represents the base station transmitting antenna gain, G r represents the receiving antenna gain of the user communication device, d BS,n represents the straight-line distance between the base station and the nth user communication device, f represents the frequency of the base station transmitting signals, 1≤n≤N, and N represents the total number of user communication devices; The signal fading model of the user communication device is expressed as: Among them, h rician represents the Rice fading channel, β is the Rice factor, which represents the strength ratio of the direct link to the equivalent multiple reflection links; h LoS represents a direct link, h NLoS Indicates an indirect link.

3. The resource allocation method based on game theory and genetic algorithm according to claim 2 is characterized in that: The S1.2 includes: Get the expression of the signal received by the user communication device from the base station: Among them, y n Indicates that the nth user communication device receives the signal from the base station, l BS,n represents the signal path loss between the base station and the nth user communication device, P n represents the signal power from the base station to the nth user communication device; x BS Indicates the signal sent by the base station, A n represents the communication frequency of the nth user communication device, A n′ A represents the communication frequency of the n′th user communication device; j represents the interference strategy of the jth interference device, that is, the interference frequency band of the jth interference device, 1≤j≤J, J represents the number of interference devices, δ(A n ,A n′ ) represents the indicator function, A n =A n′ Then δ(A n ,A n′ ) takes the value of 1, A n ≠A n′ Then δ(A n ,A n′ ) takes the value of 0; δ(A n ,A j ) represents the indicator function, A n =A j Then δ(A n ,A j ) takes the value of 1, A n ≠A j Then δ(A n ,A j ) takes the value of 0; n′ is the signal power from the base station to the n′th user communication device, l n′,n represents the signal path loss between the nth user communication device and the n′th user communication device; P j represents the signal power of the interfering device, l j,n represents the signal path loss between the interference device j and the nth user communication device, x j represents the signal emitted by the interference device, η represents the mean value is 0, and the variance is σ 2 Additive Gaussian white noise; The expression for the channel capacity from the base station to the user communication device is obtained: in, represents the channel capacity from the base station to the nth user communication device in the current resource allocation scheme, and B represents the communication bandwidth.

4. The resource allocation method based on game theory and genetic algorithm according to claim 3 is characterized in that: In step S3.3, individuals in the population are selected by using proportional selection and optimal individual retention.

5. The resource allocation method based on game theory and genetic algorithm according to claim 4 is characterized in that: In step S3.3, an adaptive crossover probability is used to perform a crossover operation on individuals in the population. The expression of the adaptive crossover probability is: Among them, p c represents the adaptive crossover probability, and Respectively represent the maximum and minimum values ​​of all adaptive crossover probabilities, f i represents the fitness of individual i, f max represents the maximum value of fitness, f avg represents the average value of fitness; The adaptive mutation probability is used to perform crossover operation on individuals in the population. The expression of the adaptive mutation probability is: Among them, p m represents the adaptive mutation probability, and They represent the maximum and minimum values ​​of the adaptive mutation probability respectively.

6. A storage medium storing a computer program for executing the steps of the resource allocation method based on game theory and genetic algorithm as claimed in any one of claims 1 to 5.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor calls the computer program in the memory, the steps of the resource allocation method based on game theory and genetic algorithm as claimed in any one of claims 1 to 5 are implemented.

Images