A dynamic on-demand cache multicast cellular network channel access method

Abstract

A dynamic on-demand cache multicast cellular network channel access method relates to the field of wireless communication network resource allocation. The small cell base station first calculates the broadcast demand arrival probability of the stored files according to the different file demand probabilities of the users, and sorts them in descending order. The base station uses the channels in the ordered channel set strategy in turn according to the current actual number of broadcast content demands. The small cell base station adjusts the current state based on the proposed dynamic channel allocation mechanism and the ordered channel set strategy of the neighboring base stations, combined with the game utility value and the parallel optimal response update strategy, and finally iteratively optimizes and converges to a stable solution. The present invention proposes a dynamic order channel allocation mechanism, combined with a potential energy game model based on minimizing interference, to achieve channel allocation optimization in a cache multicast cellular network, so as to reduce the same-frequency interference in the network, improve the channel resource utilization rate and the level of successful download of the required files of mobile users across the network.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication network resource allocation, and in particular to a dynamic on-demand cache multicast cellular network channel access method. Background Art

[0002] With the development of mobile Internet and a large number of new wireless services and applications, mobile data traffic in wireless networks has exploded, and serious problems of network backhaul congestion and redundant data transmission have also occurred. In order to improve the problems of backhaul congestion and data transmission redundancy, and further improve the utilization efficiency of channel resources, many studies have introduced local caching and multicast transmission into wireless cellular networks. Compared with traditional unicast transmission, multicast transmission can use fewer data communication channels to transmit data to multiple broadcast users within the coverage area, which can reduce the transmission of redundant data and effectively improve the utilization efficiency of wireless spectrum resources.

[0003] In most existing work, the probability of users' demand for content files follows a fixed probability distribution (Zipf distribution), and different broadcast channels are fixedly allocated for different content files. At this time, the "busy-idle" state of service transmission in the channel depends on the probability of users' demand for content files within the coverage area of ​​the base station, which will result in some channels with less co-frequency interference from adjacent base stations having relatively idle service transmission. Therefore, a dynamic broadcast channel allocation method can be designed. If there is a demand for content file broadcast transmission, idle channels with less co-frequency interference are used for transmission first. There is a lack of relevant research work on the above-mentioned dynamic channel allocation in the existing research on channel resource allocation in wireless cache multicast networks.

[0004] At the same time, due to the random distribution of the spatial locations of small cell base stations and mobile users, it is extremely complex to solve this channel combination optimization problem with a completely centralized method. Existing research often adopts a distributed self-organizing optimization method, that is, using small cell base stations as decision makers to distribute channel resources for optimal allocation. Potential game is a classic game model with high learning efficiency and fast convergence speed, and is often used to study the interaction between multiple intelligent agents. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a dynamic on-demand cached multicast cellular network channel access method, taking into account the dynamic file content requirements of users served by small cell base stations and the complex co-channel interference between neighboring base stations, a dynamic order channel allocation mechanism is proposed, and a potential energy game model based on minimizing interference is adopted to achieve channel allocation optimization in the cached multicast cellular network, so as to reduce the co-channel interference in the network, improve the channel resource utilization and the level of successful download of the required files of mobile users in the whole network.

[0006] A dynamic on-demand cache multicast cellular network channel access method comprises the following steps:

[0007] Step 1: Each base station n calculates the arrival probability of the local cache content files and arranges them in descending order to obtain the cache file set F n ={F n [1],F n [2],…,F n [C n ]}, where C n The number of local cache content files;

[0008] Step 2: Initialize the ordered channel set selection strategy of each base station and related iterative update parameters, which specifically includes the following steps:

[0009] Step 2-1: Each base station n randomly selects a C of the channel set L according to the number of content files stored locally. n The channels in the subset are randomly sorted to initialize the ordered channel allocation strategy L of base station n. n (0), and calculate the ordered channel allocation utility v of each base station n n (L n (0),L -n (0));

[0010] Step 2-2: Set the maximum value of the number of iterations K, where K is a positive integer; set the variable of the number of iterations to t, where the initial value of t is 1; set the iterative update preference probability ξ, 0<ξ<1, and the larger the ξ, the greater the probability that the base station explores a new strategy;

[0011] Step 3: Each base station n randomly selects a new strategy according to probability ξ, calculates the utility of the strategy, and updates its ordered channel set strategy according to the better response criterion; specifically, the steps include:

[0012] Step 3-1: In the tth iteration, each base station n generates a random value rand between [0,1]. During the allocation and deployment process, each base station and user in the network uses the edge cloud to transmit data, calculate utility, and coordinate decision actions in a distributed manner.

[0013] Step 3-2: If rand≤ξ, base station n attempts to update its ordered channel allocation strategy and executes step 3-3; if rand>ξ, base station n executes step 3-4;

[0014] Step 3-3: Base station n randomly selects a strategy L' that is different from the previous iteration strategy n , L' n ≠L n (t-1), calculate the utility v of the previous strategy of base station nn (L' n ,L -n (t - 1)); Compare the utility values corresponding to base station n under two different strategies: If v n (L' n ,L -n (t - 1)) > v n (L n (t - 1),L -n (t - 1)), then select the strategy L' n as the ordered channel allocation strategy for the next iteration time slot of base station n, that is, L n (t) = L' n ; If v n (L' n ,L -n (t - 1)) ≤ v n (L n (t - 1),L -n (t - 1)), then base station n keeps the current strategy unchanged, that is, L n (t) = L n (t - 1); Execute step 4;

[0015] Step 3 - 4: Base station n will keep the strategy unchanged L n (t) = L n (t - 1), and execute step 4;

[0016] Step 4: Increment the value of the iteration count variable by 1, that is, t = t + 1; If t < K, return to step 3 for execution; If t ≥ K, perform ordered channel resource allocation according to the current strategy L n (t) of the current base station.

[0017] Define the set of base stations in the small cell network as N = {1, 2,..., N}, the set of reserved wireless multicast channels as L = {1, 2,..., L}, the set of all mobile users as M = {1, 2,..., M}, and the set of all required content files as F = {1, 2,..., F}. Any content file f ∈ F has the same size and requires the same storage space. For any base station n ∈ N, C n content files have been cached (C n < F), and the set of served users is M n (the cache selection and user association of the base station have been pre - optimized and determined). For any mobile user m ∈ M, the content file demand probability follows a Zipf distribution, and the set of its content file demand probabilities is represented as P m = {p m,1 ,p m,2 ,...,p m,F}, 0 ≤ p m,f < 1,

[0018] Each base station n∈N calculates M n The joint demand arrival probability α of all users for the locally cached content file f n,f , that is, all users M served n The probability that at least one user needs to download content file f,

[0019]

[0020] The cache file set sorted in descending order of file demand probability in base station n∈N is denoted as F n ={F n [1],F n [2],…,F n [C n ]}, F n [k] is the kth content file in the cache file set. For any two files in the set, n [k] and F n [k'], if F n [k] Sort relative to F n [k'] is at the front, i.e. 0 <k<k'≤C n ,but The ordered channel set strategy selected by base station n is represented by L n ={L n [1],L n [2],…,L n [C n ]},

[0021] Base station n prefers to use the top ranked channels according to the number of files that need to be broadcast. After iterative optimization of channel selection, the channels with smaller co-channel interference are ranked higher. For base station n, it is possible to use channels with smaller co-channel interference with the greatest possible probability, ultimately reducing the co-channel interference level of the overall network and improving the overall download success rate of network multicast files. For each base station n, the set of broadcast demand files that arrive at a certain time is actually its cache file set F n A subset of (not considering the channel allocation of uncached demand files). Indicates F n The set of all k-ary subsets of , the overall F n The power set of can be expressed as The probability that base station n receives broadcast requests for k content files at a certain moment, that is, F n The probability of a k-ary subset of can be expressed as

[0022]

[0023] At this time, base station n only needs to use the ordered channel set strategy L n The first k channels in the system are used for content file broadcast transmission. In this case, the set of transmission channels used is represented by L n [1:k]={L n [1],L n [2],…,L n [k]}. It can also be inferred that under the proposed dynamic channel allocation mechanism, when the first k channels are needed for content file broadcast transmission, the hth (k>h)th channel in the ordered channel set strategy will also be occupied. Therefore, base station n uses L n The hth channel L n [h](L n [h],0 <h≤C n ) can be expressed as

[0024]

[0025] Since the base station cache files are sorted in descending order according to the demand arrival probability, the broadcast channel allocation is carried out in an orderly manner, so F n The kth content file F n The channel allocation of [k] is only affected by the first k-1 content files (whose set is represented by F n [1:k-1]). A user m∈M of base station n n Through channel L n [h] Download content file F n The probability of [k] can be expressed as

[0026]

[0027] in, For user m, content file F n [k] demand probability. When k = 1, h = 1, and Since the base station transmission power is fixed, its interference range is limited. The set of all users within the interference range of base station n is expressed as The set of potentially interfering base stations for user m is denoted as φ m ; Then user m downloads content file F from base station n n [k] is {φ m The probability of interference from any base station in

[0028]

[0029] Among them, n′∈{φ m -n} represents any base station in the set of potential interfering base stations of user m except base station n.

[0030] The overall interference level experienced by user m can be expressed as

[0031]

[0032] The overall interference level of the entire network can be expressed as

[0033] In the process of optimizing the ordered channel allocation of base stations, the base station minimizes the sum of the interference levels of all users within its interference range by selecting different ordered channel set strategies, while achieving a continuous decrease in the overall co-frequency interference level of the entire network. The ordered channel allocation utility design of base station n in this stage is as follows:

[0034]

[0035] Where L -n represents the ordered channel allocation strategy of other base stations except base station n. At this time, the potential energy function can be constructed This is the overall interference level of the entire network. When any base station n unilaterally changes its strategy and increases the corresponding utility v n When the potential energy function Φ n The value of will also increase synchronously, that is, the global interference level decreases.

[0036] The beneficial effects of the present invention are:

[0037] 1. The present invention proposes a dynamic order channel allocation mechanism for wireless multicast networks. In the proposed dynamic channel allocation mechanism, the channel allocation strategy of the base station is an ordered channel set. The base station uses the channels in the ordered channel set strategy in turn according to the actual number of broadcast content requirements. Compared with the traditional fixed channel allocation mechanism for broadcast content, the proposed dynamic order channel allocation mechanism can maximize the use of channels with less co-frequency interference, while reducing the overall interference level of the entire network and improving the channel utilization efficiency.

[0038] 2. The present invention adopts a potential game model and a distributed decision-making architecture to solve this complex non-convex discrete nonlinear programming problem, and constructs a corresponding utility function for the channel allocation strategy of the base station. The base station can optimize the global utility while optimizing its own utility.

[0039] 3. The present invention adopts a multi-base station parallel channel allocation algorithm suitable for the proposed dynamic order channel allocation mechanism and potential energy game. It can change the exploration and utilization probability of the learning algorithm through reasonable parameter settings, can adapt to a variety of complex network environments, and has great application potential in actual wireless network resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1This is a diagram of the wireless multicast network model in the present invention.

[0041] Figure 2 It is a schematic diagram of the dynamic ordered channel allocation mechanism proposed by the present invention.

[0042] Figure 3 It is a working flow chart of the network channel access method of the present invention.

[0043] Figure 4 It is a comparison chart of the overall network interference level between the network channel access method of the present invention and two channel allocation methods when the number of local cache files in the base station changes.

[0044] Figure 5 It is a comparison diagram of the overall network interference level of the network channel access method of the present invention and two channel allocation methods when the number of reserved channels changes.

[0045] Figure 6 It is a comparison chart of the successful downloading level of the whole network file when the number of reserved channels changes between the network channel access method of the present invention and two channel allocation methods. DETAILED DESCRIPTION

[0046] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0047] The small cell base station first calculates the broadcast demand arrival probability of the stored files according to the different file demand probabilities of the users, and sorts them in descending order. Different from the "content--channel" fixed allocation mechanism of traditional multicast transmission, the channel allocation strategy of the small cell base station in the proposed dynamic channel allocation mechanism is an ordered channel set. The base station uses the channels in the ordered channel set strategy in turn according to the current actual number of broadcast content demands. Specifically, if the base station currently has k files to be broadcast (determined by the user demand probability), according to the sorting of the above cached files, the first k channels in the ordered channel set strategy are occupied in turn from front to back according to demand. On this basis, the small cell base station uses a game model that minimizes interference potential energy to calculate the game utility value of the corresponding strategy based on the proposed dynamic channel allocation mechanism and the ordered channel set strategy of neighboring base stations, and then adjusts the current state according to a parallel optimal response update strategy, and finally iterates and optimizes to converge to a stable solution.

[0048] like Figure 1As shown in the figure, the cached multicast cellular network consists of N cellular base stations with local caching capabilities (the set of base stations is denoted as N = {1, 2, …, N}) and M mobile users (the set of users is denoted as M = {1, 2, …, M}). All mobile users have a total demand for F content files (the F content files are of the same size, require the same storage space, and can be packaged through a data gateway. The file set is denoted as F = {1, 2, …, F}). For any base station n ∈ N, C n content files have been cached (C n < F), and the set of served users is M n (the cache selection and user association of the base station have been pre-optimized and determined). The content file demand probability of any mobile user m ∈ M follows a Zipf distribution, and the set of its demand probabilities for content files is denoted as P m = {p m,1 , p m,2 , …, p m,F}, 0 ≤ p m,f < 1, where π(f, m) is the preference ranking number of the user group of the same type as mobile user m for the content file, and γ > 0 is the preference parameter).

[0049] According to the transmission power of the small cell base station, its transmission coverage range and communication interference range can be determined. Generally speaking, the distance of communication interference is twice the transmission distance. When different small cell base stations use the same channel for multicast transmission, co-channel interference may occur. At this time, the interfered mobile users will not be able to successfully download the corresponding content files. Therefore, it is necessary to optimize the allocation of channel resources in the cached multicast cellular network to reduce co-channel interference (co-frequency interference) during the multicast transmission of content files and improve the successful download level of the content files demanded by mobile users across the network. The set of all users within the interference range of base station n is denoted as The set of potential interfering base stations of user m is denoted as φ m . As Figure 1 shown in the figure, for small cell base station 1, the users it serves are users 1, 2, 3, 7, and the users within the interference range of small cell base station 1 (φ1) are 1, 2, 3, 5, 7; for user 5, its potential interfering base stations are base station 1 and base station 2.

[0050] Next, taking Figure 2 as an example, the proposed dynamic order channel allocation mechanism will be specifically described.

[0051] First, each base station n ∈ N calculates the joint demand arrival probability α n of all users in M n,f for the locally cached content file f, that is, all served users M nThe probability that at least one user needs to download content file f,

[0052]

[0053] Each base station sorts the files in descending order according to the broadcast demand arrival probability of the stored files. The cache file set sorted in descending order according to the file demand probability in base station n∈N is represented as F n ={F n [1],F n [2],…,F n [C n ]}, F n [k] is the kth content file in the cache file set. For any two files in the set, n [k] and F n [k'], if F n [k] Sort relative to F n [k'] is at the front, i.e. 0 <k<k'≤C n ,but like Figure 2 As shown, after calculating the broadcast demand arrival probability of each stored file, the ordered set F1 of files stored in base station 1 is {file 1, file 3, file 4, file 5, file 2}.

[0054] Then, base station n selects an ordered channel allocation strategy L n ={L n [1],L n [2],…,L n [C n ]}, Base station n prefers to use the channels with the highest ranking according to the number of files to be broadcast. Figure 2 As shown, the ordered channel set strategy L1 allocated by base station 1 is {channel 2, channel 3, channel 6, channel 1, channel 7}. Base station 1 gives priority to using the channels with the highest ranking according to the number of required broadcast files that actually arrive. At the first moment, base station 1 currently has 3 required broadcast files (files 1, 4, and 2). According to the ranking of the cached files, the first 3 channels in the ordered channel set strategy are occupied from front to back as needed (file 1 uses channel 2, file 4 uses channel 3, and file 2 uses channel 6); at the second moment, base station 1 currently has 2 required broadcast files (file 3 and file 5). According to the ranking of the cached files, L1 is occupied from front to back as needed. nThe first two channels in (file 3 uses channel 2, and file 5 uses channel 3). Through subsequent continuous iterative optimization, for a certain base station, it can give higher priority to channels with less co-channel interference (i.e., rank them first), and use channels with less co-channel interference with the greatest probability, ultimately reducing the co-channel interference level of the entire network and improving the overall download success level of network multicast files.

[0055] like Figure 3 As shown, the proposed method for dynamic allocation of wireless channels in a cached multicast cellular network first calculates the broadcast demand arrival probability of the stored files according to the different file demand probabilities of the users, and sorts them in descending order; based on the proposed dynamic channel allocation mechanism, the small cell base station uses a game model that minimizes interference potential energy to calculate the game utility value of the corresponding strategy based on the proposed dynamic channel allocation mechanism and the ordered channel set strategy of the neighboring base stations, and then adjusts the current state according to a parallel optimal response update strategy, and finally iterates and optimizes to converge to a stable solution. Further, step 3 The game utility calculation and strategy update process is as follows:

[0056] For each base station n, the broadcast demand file set arriving at a certain time is actually the cache file set F n A subset of (not considering the channel allocation of uncached demand files). Indicates F n The set of all k-ary subsets of , then F n The power set of can be expressed as The probability that base station n receives broadcast requests for k content files at a certain moment, that is, F n The probability of a k-ary subset of can be expressed as

[0057]

[0058] At this time, base station n only needs to use the ordered channel set strategy L n The first k channels in the system are used for content file broadcast transmission. In this case, the set of transmission channels used is represented by L n [1:k]={L n [1],L n [2],…,L n [k]}. It can also be inferred that under the proposed dynamic channel allocation mechanism, when the first k channels are needed for content file broadcast transmission, the hth (k>h)th channel in the ordered channel set strategy will also be occupied. Therefore, base station n uses L n The hth channel L n [h](L n [h],0 <h≤C n ) can be expressed as

[0059]

[0060] Since the base station cache files are sorted in descending order according to the demand arrival probability, the broadcast channel allocation is carried out in an orderly manner, so F n The kth content file F n The channel allocation of [k] is only affected by the first k-1 content files (whose set is represented by F n [1:k-1]). A user m∈M of base station n n Through channel L n [h] Download content file F n The probability of [k] can be expressed as

[0061]

[0062] In particular, when k = 1, h = 1, and Since the base station transmission power is fixed, its interference range is limited. The set of all users within the interference range of base station n is expressed as The set of potentially interfering base stations for user m is denoted as φ m Then user m downloads content file F from base station n n [k] is {φ m The probability of base station interference in the -n} can be expressed as

[0063]

[0064] The overall interference level experienced by user m can be expressed as

[0065]

[0066] The overall interference level of the entire network can be expressed as

[0067] In the process of optimizing the ordered channel allocation of base stations, the base station minimizes the sum of the interference levels of all users within its interference range by selecting different ordered channel set strategies, while achieving a continuous decrease in the overall co-frequency interference level of the entire network. The ordered channel allocation utility design of base station n in this stage is as follows:

[0068]

[0069] Where L -n represents the ordered channel allocation strategy of other base stations except base station n. At this time, the potential energy function can be constructed This is the overall interference level of the entire network. When any base station n unilaterally changes its strategy and increases the corresponding utility v n When the potential energy function Φ n The value of will also increase synchronously, that is, the global interference level decreases.

[0070] The present invention adopts Figure 3 The process shown includes the following steps:

[0071] Step 1: Each base station n calculates the arrival probability of the local cache content file according to formula (1) and arranges them in descending order to obtain the arrangement set F n ={F n [1],F n [2],…,F n [C n ]}.

[0072] Step 2: Initialize the ordered channel set selection strategy of each base station and related iterative update parameters, which specifically includes the following steps:

[0073] Step 2-1: Each base station n randomly selects a C of the channel set L according to the number of content files stored locally. n The channels in the subset are randomly sorted to form a C n Meta-ordered channel set strategy L n (0), that is, initialize the ordered channel allocation strategy of each base station n, and calculate the ordered channel allocation utility v of each base station n n (L n (0),L -n (0)).

[0074] Step 2-2: Set the maximum number of iterations K, where K is a positive integer (the more base stations there are in the network, the more cache space and the more reserved broadcast channels there are, the larger the K value should be); set the variable of the number of iterations to t, and the initial value of t is 1; set the iteration process preference factor ξ (0<ξ<1, which affects the frequency of occurrence of small cell base station and mobile user policy update processes in the iteration).

[0075] Step 3: Each base station n randomly selects a new strategy according to probability ξ, calculates the utility of the strategy, and updates its ordered channel set strategy according to the better response criterion. Step 3 specifically includes the following steps:

[0076] Step 3-1: In the tth iteration, each base station n generates a random value rand between [0,1]. During the allocation and deployment process, each base station and user in the network uses edge cloud technology to transmit relevant data, calculate utility, and coordinate decision actions in a distributed manner.

[0077] Step 3-2: If rand≤ξ, base station n attempts to update its ordered channel allocation strategy and executes step 3-3; if rand>ξ, base station n executes step 3-4.

[0078] Step 3-3: Base station n randomly selects a strategy L' that is different from the previous iteration strategy n (L' n ≠L n (t - 1)), and calculates the utility v of base station n under the previous strategy according to formula (7) n (L' n ,L -n (t - 1)). Compare the magnitudes of the utility values of base station n under the two different strategies: If v n (L' n ,L -n (t - 1)) > v n (L n (t - 1),L -n (t - 1)), then select strategy L' n as the ordered channel allocation strategy for the next iteration time slot of base station n, that is, L n (t) = L' n ; If v n (L' n ,L -n (t - 1)) ≤ v n (L n (t - 1),L -n (t - 1)), then base station n keeps the current strategy unchanged, that is, L n (t) = L n (t - 1). Execute Step 4.

[0079] Step 3-4: Base station n keeps the strategy unchanged L n (t) = L n (t - 1), and executes Step 4.

[0080] Step 4: Increment the value of the iteration count variable by 1, that is, t = t + 1. If t < K, return to Step 3 for execution; if t ≥ K, perform ordered channel resource allocation according to the current strategy L n (t) of the current base station.

[0081] The present invention will be further described below in conjunction with specific embodiments.

[0082] The effectiveness of the present invention is verified by simulation examples. First, the scenario of the embodiment is briefly introduced. Five small cell base stations are randomly distributed in an area with a radius of 50m. The transmission distance of the base station is 15m, the interference distance is set to 30m, each small cell base station serves 11 to 13 mobile users (mobile users are randomly distributed within the effective transmission range of the small cell base station), the total number of content files is F=10, and the demand of mobile users for content files obeys the Zipf distribution (γ=0.8). Each base station pre-caches the C most popular content files, the maximum number of iterations is 1000 times, the number of reserved channels is not less than the number of files cached by the base station (L≥C), and the parameter setting does not affect the generality.

[0083] Then, three wireless multicast network resource allocation methods were run using the Matlab simulation platform: ① the dynamic order channel allocation algorithm (ξ=0.8) proposed in this invention; ② the fixed channel allocation algorithm based on the optimal response; ③ random channel resource allocation (fixed channel allocation). The relevant comparison results are shown in Figure 2. Figure 4 , Figure 5 and Figure 6 As shown, it can be found that the dynamic channel allocation method proposed in the present invention achieves the best effect. Compared with random allocation and fixed channel allocation algorithm based on better response, it can significantly reduce the overall interference level of the entire network and improve the successful download level of file content for users in the entire network.

[0084] Figure 4 It is the overall interference level of the whole network when the number of cache files (cache space) of the base station changes. Among them, the number of reserved channels is L=10. From the analysis of this figure, it can be seen that the overall co-channel interference level of the network proposed by the algorithm of the present invention is lower than that of other methods. At the same time, as the cache space of the base station increases, the number of channels required for content broadcasting increases, and the overall co-channel interference level of the network increases accordingly.

[0085] Figure 5 and Figure 6 They are the overall interference level of the entire network and the level of successful file downloads when the number of reserved channels changes. Among them, the cache space of all base stations is C = 8. Analysis of this figure shows that as the number of reserved channels increases, the overall interference level of the network decreases (after decreasing to a certain level, the decreasing speed slows down), and the level of successful file downloads of users in the entire network also increases (after increasing to a certain level, the growth rate slows down).

[0086] The present invention has been described above in an illustrative manner using embodiments. Those skilled in the art should understand that the present disclosure is not limited to the embodiments described above, and that various changes, modifications and substitutions may be made without departing from the scope of the present invention, and that the parts not covered by the present invention are the same as the prior art or may be implemented using the prior art.

Claims

1. A dynamic on-demand cache multicast cellular network channel access method, characterized in that The following steps are involved: Step 1: Each base station n calculates the arrival probability of the local cache content files and arranges them in descending order to obtain the cache file set F n ={F n [1],F n [2],…,F n [C n ]}, where C n The number of local cache content files; Step 2: Initialize the ordered channel set selection strategy of each base station and related iterative update parameters, which specifically includes the following steps: Step 2-1: Each base station n randomly selects a C of the channel set L according to the number of content files stored locally. n The channels in the subset are randomly sorted to initialize the ordered channel set strategy L of base station n. n (0), and calculate the ordered channel allocation utility v of each base station n n (L n (0),L -n (0)); Step 2-2: Set the maximum value of the number of iterations K, where K is a positive integer; set the variable of the number of iterations to t, where the initial value of t is 1; set the iterative update preference probability ξ, 0<ξ<1, and the larger the ξ, the greater the probability that the base station explores a new strategy; Step 3: Each base station n randomly selects a new strategy according to probability ξ, calculates the utility of the strategy, and updates its ordered channel set strategy according to the better response criterion; specifically, the steps include: Step 3-1: In the tth iteration, each base station n generates a random value rand between [0,1]. During the allocation and deployment process, each base station and user in the network uses the edge cloud to transmit data, calculate utility, and coordinate decision actions in a distributed manner. Step 3-2: If rand≤ξ, base station n attempts to update its ordered channel allocation strategy and executes step 3-3; if rand>ξ, base station n executes step 3-4; Step 3-3: Base station n randomly selects a strategy L' that is different from the previous iteration strategy n , L' n ≠L n (t-1), calculate the utility v of the previous strategy of base station n n (L' n ,L -n (t-1)); compare the utility values ​​of base station n under two different strategies: if v n (L' n ,L -n (t-1))>v n (L n (t-1),L -n (t-1)), then choose strategy L' n As the ordered channel allocation strategy for the next iteration time slot of base station n, that is, L n (t) = L' n ; if v n (L' n ,L -n (t-1))≤v n (L n (t-1),L -n (t-1)), then base station n maintains the current strategy unchanged, that is, L n (t) = L n (t-1); execute step 4; Step 3-4: Base station n will keep the strategy unchanged L n (t) = L n (t-1), execute step 4; Step 4: Increment the value of the iteration number variable by 1, i.e., t = t + 1; if t < K, return to step 3 for execution; if t ≥ K, perform ordered channel resource allocation according to the policy L n (t).

2. The method for dynamic on-demand cache multicast cellular network channel access according to claim 1, characterized in that The specific process of the above step 1 is as follows: Define the set of base stations in the small cell network as N = {1, 2, …, N}, the set of reserved wireless multicast channels as L = {1, 2, …, L}, the set of all mobile users as M = {1, 2, …, M}, and the set of all required content files as F = {1, 2, …, F}. Any content file f ∈ F has the same size and requires the same storage space. For any base station n ∈ N, C n content files have been cached, C n < F, and the set of served users is M n ; For any mobile user m ∈ M, the content file demand probability follows the Zipf distribution, and the set of content file demand probabilities is expressed as P m = {p m,1 , p m,2 , …, p m,F}, 0 ≤ p m,f < 1, Each base station n∈N calculates M n The joint demand arrival probability α of all users for the locally cached content file f n,f , that is, all users M served n The probability that at least one user needs to download content file f, The cache file set sorted in descending order of file demand probability in base station n∈N is denoted as F n ={F n [1],F n [2],…,F n [C n ]}, F n [k] is the kth content file in the cache file set. For any two files in the set, n [k] and F n [k'], if F n [k] Sort relative to F n [k'] is at the front, i.e. 0 <k<k'≤C n ,but The ordered channel set strategy selected by base station n is represented by L n ={L n [1],L n [2],…,L n [C n ]}, 3. The method for dynamic on-demand cache multicast cellular network channel access according to claim 2, characterized in that The specific process of step 2 above is: For each base station n, the set of broadcast demand files arriving at a certain time is actually its cache file set F n A subset of Represents the cache file set F n The set of all k-ary subsets of the overall cache file set F n The power set of The probability that base station n receives a total of k content files’ broadcast requests at a certain moment, i.e., F n The probability of a k-ary subset of At this time, base station n only needs to use the ordered channel set strategy L n The first k channels in the transmission channel are used for content file broadcast transmission. The transmission channel set used at this time is represented by L n [1:k]={L n [1],L n [2],…,L n [k]}; It can be inferred that under the proposed dynamic channel allocation mechanism, when the first k channels are needed for content file broadcast transmission, the hth channel in the ordered channel set strategy will also be occupied, k>h; therefore, base station n uses L n The hth channel L n The probability of [h] is expressed as: Where 0 <h≤C n , since the base station cache files are sorted in descending order according to the demand arrival probability, the broadcast channel allocation is carried out in an orderly manner, then F n The kth content file F n The channel allocation of [k] is only affected by the first k-1 content files, and the set is represented by F n [1:k-1]; A user m∈M of base station n n Through channel L n [h] Download content file F n The probability of [k] is expressed as: in, For user m, content file F n [k] demand probability; when k = 1, h = 1, and Since the base station transmission power is fixed, its interference range is limited. The set of all users within the interference range of base station n is expressed as The set of potentially interfering base stations for user m is denoted as φ m ; Then user m downloads content file F from base station n n [k] is {φ m The probability of interference from any base station in Among them, n′∈{φ m -n} represents any base station in the set of potential interfering base stations for user m except base station n; The overall interference level experienced by user m is expressed as The overall interference level of the whole network is To express.

4. The method for dynamic on-demand cache multicast cellular network channel access according to claim 1, characterized in that The specific process of step 3 above is: In the process of optimizing the ordered channel allocation of base stations, the base station minimizes the sum of the interference levels of all users within its interference range by selecting different ordered channel set strategies, while achieving a continuous decrease in the overall co-frequency interference level of the entire network. The ordered channel allocation utility design of base station n in this stage is as follows: Where L -n represents the ordered channel allocation strategy of other base stations except base station n; At this point, the potential energy function can be constructed is the overall interference level of the entire network; when any base station n unilaterally changes its strategy and increases the corresponding utility v n When the potential energy function Φ n The value of will also increase synchronously, that is, the global interference level decreases.

Images