Time slot allocation method and system based on adaptive weight particle swarm optimization algorithm

Abstract

The invention provides a time slot allocation method and system based on an adaptive weight particle swarm optimization algorithm, and the method comprises the steps: 1, determining service parameters according to the service demands of all network nodes in a data link network; 2, initializing algorithm parameters; step 3, calculating time delay jitter when a single node allocates time slot resources; step 4, calculating an average value of delay jitter of each node when the multi-node time slot resource is allocated; step 5, updating the optimal time slot pbest and the global optimal time slot gbest of the network node; 6, updating the inertia weight W according to an adaptive weight particle swarm optimization algorithm; and 7, updating the speed and the position of the particle according to the inertia weight updated in the step 6. According to the method, the situation that the particle swarm optimization is prone to falling into a local optimal solution due to the fact that the inertia weight is too large or too small in iteration can be avoided, the convergence speed of the last stage of iteration is increased, the algorithm performance is improved, and the optimal time slot allocation scheme of the data chain can be quickly and accurately found.

Description

Technical Field

[0001] The present application relates to the technical field of data link time slot allocation, and in particular to a time slot allocation method and system based on an adaptive weighted particle swarm optimization algorithm. Background Art

[0002] Time Division Multiple Access (TDMA) has become a very common multi-access network communication method. Tactical data link is a wireless communication system used in the military field. It uses TDMA for networking and has the characteristics of efficient transmission, high confidentiality requirements and strong anti-interference ability. The data link can perform many tasks such as radio communication, air surveillance, aviation combat command and control, and realize seamless exchange of messages between air, ground and marine forces during joint operations. The application of tactical data link has an increasingly greater impact on the modern information warfare situation, and many countries have also invested in the research and development of data links.

[0003] The TDMA access method of the data link network divides time resources into time units, time frames and time slots. One time unit can contain M time frames, and one time frame contains N time slots. The time slot is the basic time unit of the data link network. The time slot resources are allocated to each node member in the network through the time slot allocation algorithm, and each node transmits information in the allocated time slot.

[0004] When allocating time slot resources, in order to enable the periodic transmission of tactical messages in the data link, it is necessary to reduce the delay of the message during transmission, so as to ensure the reliability and stability of the tactical data link. When successfully allocating time slots to network nodes, it is necessary to ensure the uniformity of the allocated time slots within a time frame as much as possible. The more uniform the time slot allocation, the shorter the delay in the transmission of business messages will be, and the smaller the delay jitter will be. Many scholars at home and abroad have put forward their many insights on the method of allocating time slots for tactical data links.

[0005] a. Time slot allocation algorithm based on binary tree time slot block

[0006] At present, most of the data link time slot allocation algorithms start from the uniformity of time slot allocation. The most classic time slot allocation algorithm based on binary tree time slot blocks can only guarantee the uniformity of time slot allocation within a time frame, but cannot guarantee that the time slots allocated to periodic messages in another time frame are also uniform. Since the time slot blocks based on binary trees are composed of time slots that are powers of 2, and in practice, the number of time slots required for business messages is mostly not a power of 2. Therefore, using binary tree time slot blocks to complete time slot resource allocation will result in more time slots not being used, resulting in a waste of time slot resources.

[0007] b. Average domain annealing algorithm

[0008] In IEEE Journal on selected areas in Communications, 1997, 15(2): 250-260, Wang G et al. published a paper titled "Optimal broadcast scheduling in packet radio networks using mean field annealing". In this paper, a mean field annealing algorithm was disclosed. However, since the algorithm has a large amount of computation, it takes a lot of time to obtain a reasonable time slot allocation scheme, which does not meet the real-time nature of information transmission in tactical data links.

[0009] c. Time slot allocation algorithm based on hybrid neural genetic algorithm

[0010] In IEEE Transactions on Wireless Communications,2003,2(2):

[0011] In 277-283, Salcedo-Sanz S et al. published a paper titled "A mixed neural-genetical algorithm for the broadcast scheduling problem", in which they disclosed a time slot allocation algorithm based on a mixed neural-genetic algorithm. However, the algorithm is relatively complex to implement and is not suitable for practical engineering use.

[0012] d. TDMA time slot allocation algorithm based on genetic algorithm

[0013] In Firepower and Command and Control, 2015, 40(08):97-100, Chen Jiayuan et al. published a paper titled "Time Slot Allocation Algorithm for TDMA Tactical Data Link Based on Genetic Algorithm". The paper proposed a TDMA time slot allocation algorithm based on genetic algorithm. The algorithm breaks the constraints of the binary tree structure and directly searches for the optimal solution from the solution space of time slot allocation through the traditional genetic algorithm. However, the genetic algorithm has poor local search capabilities, and the training time required to obtain a reliable solution is too long, which cannot meet the real-time requirements of information transmission of tactical data links during combat.

[0014] e. Data link network resource allocation method based on genetic algorithm

[0015] In IEEE, 2019:1875-1880, Sun W et al. published the paper "Data link network resource allocation method based on genetic algorithm", which proposed a data link network resource allocation method based on genetic algorithm. Due to the poor local search ability of genetic algorithm, it will take too long to train to obtain the optimal solution of reliable time slots, so it cannot meet the real-time requirements of tactical data link for information transmission during combat.

[0016] f. A time division multiple access time slot allocation method based on manta ray algorithm

[0017] The patent application with publication number CN116234017A provides a time division multiple access time slot allocation method based on the manta ray algorithm, which divides the time frame into an information collection period and a time slot allocation period, and then divides the time slot allocation period into a dedicated time slot period and a dynamic time slot period. However, when the time division multiple access time slot allocation method based on the manta ray algorithm is used to obtain the optimal time slot resource allocation scheme, its delay jitter is too large and the theoretical optimal situation is not reached.

[0018] In summary, the existing time slot allocation technology will cause more time slots to be unused, resulting in a waste of time slot resources; the algorithm has a large amount of computation, and it takes a lot of time to obtain a reasonable time slot allocation plan, which does not meet the real-time information transmission of the tactical data link; the algorithm implementation is relatively complex and is not suitable for actual engineering use; and the algorithm has poor local search capabilities. In order to obtain a reliable solution, the training time is too long, which cannot meet the real-time requirements of data link information transmission during combat and has large delay jitter. Summary of the invention

[0019] In view of this, the present application provides a time slot allocation method and system based on an adaptive weighted particle swarm optimization algorithm to reduce delay jitter and improve the combat effectiveness of tactical data links and the stability of information transmission.

[0020] To achieve the above objectives, the technical solutions adopted in this application are as follows:

[0021] In the first aspect, the present application provides a time slot allocation method based on an adaptive weighted particle swarm optimization algorithm, which takes a single network node in a data link network as a particle, and obtains an overall optimal time slot allocation scheme for the network node by continuously moving the particle toward the optimal solution direction, and comprises the following steps:

[0022] Step 1: Determine service parameters according to the service requirements of each network node in the data link network, wherein the service parameters include: the number of nodes with time slot requirements m, the total number of idle time slots S and the time slot resource requirement n;

[0023] Step 2: Initialize the algorithm parameters, including: particle swarm size M, particle position X, particle velocity V, learning factors C1 and C2, inertia weight range [W min ,W max ], the range of particle flight speed [v min ,v max ] and the maximum number of iterations K;

[0024] Step 3: Determine the interval between adjacent time slots of a single node Is the difference between the adjacent time slot interval P of the theoretical optimal time slot allocation less than or equal to the minimum value min? If so, jump to step 8, otherwise calculate the delay jitter when a single node allocates time slot resources;

[0025] Step 4: Calculate the average value of the delay jitter of each node when multiple nodes allocate time slot resources;

[0026] Step 5: Update the optimal time slot p of the network node according to the delay jitter results calculated in steps 3 and 4 best and the global optimal time slot g best ;

[0027] Step 6: Update the inertia weight W according to the adaptive weight particle swarm optimization algorithm;

[0028] Step 7: Update the particle speed and position according to the inertia weight updated in step 6;

[0029] Step 8: Repeat steps 3 to 7, and increase the number of iterations by 1. When the number of iterations reaches the maximum number of iterations, stop the iteration, and use the optimal position of each particle as the optimal time slot allocation scheme for each network node and output it.

[0030] Furthermore, the specific method for calculating the delay jitter when a single node allocates time slot resources in step 3 is:

[0031]

[0032] Said is the interval between the i-th adjacent time slots allocated to node j in the k-th iteration, nj is the time slot resource requirement of the j-th node, and from formula (1), P is the total idle time slots S in a frame divided by the number of time slots n that node j needs to allocate j , that is, P is the adjacent time slot interval of the theoretically optimal time slot allocation.

[0033] Furthermore, the formula for calculating the average value of the delay jitter of each node when multiple nodes allocate time slot resources in step 4 is: fit = min {fit k ,k=1,2,...K}

[0034] in

[0035] The fit is a fitness function, which represents the average delay jitter of each node, and m is the number of nodes with time slot requirements in the data link network. The delay jitter generated when allocating time slots to the above node j.

[0036] Furthermore, in step 5, the optimal time slot p of the node is obtained. best and the global optimal time slot g best The method is:

[0037] For each particle, the delay jitter result calculated by formula (3) is compared with the result of the last delay jitter of the particle, and the particle position corresponding to the smaller delay jitter is taken as the current optimal position of the particle, and the optimal time slot p of the node is updated accordingly. best ;

[0038] Compare the average delay jitter result output by formula (5) with the average delay jitter value of the previous iteration, and take the smaller average delay jitter as the value p best As the global optimal position, and update the global optimal time slot g accordingly best .

[0039] Furthermore, the updating formula for the inertia weight in step 6 is as follows:

[0040] W=W max -(W max -W min )×(k / K) 2 (6)

[0041] In formula (6), W is the inertia weight, W max , W min They respectively represent the maximum and minimum values ​​of the inertia weight W, where k represents the number of iterations of the current program and K represents the maximum number of iterations.

[0042] Furthermore, in step 7, the velocity update formula and position update formula for the particle are (7) and (8) respectively:

[0043]

[0044] Where:

[0045] k is the number of iterations;

[0046] W is the inertia weight updated in step 5;

[0047] p i represents the optimal position of a single particle, which is equivalent to p best ;pg represents the optimal position of the global particle, which is equivalent to g best ;

[0048] C1 and C2 are learning factors, where C1 is the individual learning factor of each particle and C2 is the individual social learning factor of each particle; r1 and r2 are random numbers between 0 and 1.

[0049] In the second aspect, the present application provides a time slot allocation system based on an adaptive weighted particle swarm optimization algorithm, which uses a single network node in a data link network as a particle, and continuously moves the particle toward the optimal solution direction to obtain an overall optimal time slot allocation scheme for the network node, including:

[0050] A network initialization module, used to determine service parameters according to the service requirements of each network node in the data link network, wherein the service parameters include: the number of nodes with time slot requirements m, the total number of idle time slots S and the time slot resource requirement n;

[0051] The particle swarm initialization module is used to initialize the algorithm parameters, including: particle swarm size M, particle position X, particle velocity V, learning factors C1 and C2, inertia weight range [W min ,W max ], the range of particle flight speed [v min ,v max ] and the maximum number of iterations K;

[0052] The first calculation module is used to calculate the delay jitter when a single node allocates time slot resources;

[0053] The second calculation module is used to calculate the average value of the delay jitter of each node when multiple nodes allocate time slot resources;

[0054] A first updating module is used to update the optimal time slot p of the network node according to the average value of the delay jitter calculated by the first calculation module and the delay jitter calculated by the second calculation module. best and the global optimal time slot g best ;

[0055] The second updating module is used to update the inertia weight W according to the adaptive weight particle swarm optimization algorithm;

[0056] A third updating module, used for updating the particle speed and position according to the inertia weight updated by the second updating module;

[0057] The optimal time slot allocation module is used to repeatedly execute steps 3 to 7, and increase the number of iterations by 1. When the number of iterations reaches the maximum number of iterations, the iteration is stopped, and the optimal position of each particle is used as the optimal time slot allocation scheme of each network node and outputted.

[0058] Furthermore, the second updating module is specifically used to update the inertia weight according to formula (6):

[0059] W=W max -(W max -W min )×(k / K) 2 (6)

[0060] In formula (6), W is the inertia weight, W max , W min They respectively represent the maximum and minimum values ​​of the inertia weight W, where k represents the number of iterations of the current program and K represents the maximum number of iterations.

[0061] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the time slot allocation method based on the adaptive weighted particle swarm optimization algorithm as described in the first aspect are implemented.

[0062] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the time slot allocation method based on the adaptive weighted particle swarm optimization algorithm as described in the first aspect are implemented.

[0063] The beneficial effects of this application are:

[0064] Compared with the prior art, the beneficial effects of this application are:

[0065] 1. The adaptive inertia weight update strategy based on the particle swarm optimization algorithm (PSO) adopted in this application can prevent the particle swarm algorithm from falling into the local optimal solution due to the inertia weight being too large or too small during iteration, accelerate the convergence speed at the end of iteration, improve the algorithm performance, and can quickly and accurately find the optimal time slot allocation scheme for the data link, and can significantly reduce the delay jitter, and quickly and accurately find the optimal solution for time slot allocation.

[0066] 2. This application uses delay jitter to evaluate the uniformity of time slot allocation. A smaller interval represents time slot allocation that is closer to the adjacent time slot interval of the theoretically optimal time slot allocation. From the simulation data, it can be seen that the iterative convergence speed of the time slot allocation method is faster, so that the time slot allocation method can achieve maximum performance, that is, it can find the optimal time slot allocation plan faster and more efficiently.

[0067] 3. This application adopts a data link time slot allocation method based on an adaptive weighted particle swarm optimization algorithm to complete time slot resource allocation. Compared with the prior art, it can better help network planners to quickly make corresponding time slot allocation plan plans based on combat mission conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0069] Figure 1 A flowchart of the time slot allocation method for adaptive weighted particle swarm optimization (AWPSO-SAA) provided in this application;

[0070] Figure 2 It is the relationship between the number of iterations and delay jitter during single-node time slot allocation;

[0071] Figure 3 The relationship between the number of iterations and delay jitter during multi-node time slot allocation;

[0072] Figure 4 The fitness value calculation process of each node in each generation.

[0073] Figure 5 The time complexity of AWPSO-SAA algorithm for this application;

[0074] Figure 6 Validation diagram for the stability of the time slot allocation algorithm;

[0075] Figure 7 It is the correlation diagram between the number of particle groups and time delay;

[0076] Figure 8 It is the correlation diagram between the number of idle time slots and delay;

[0077] Fig. 9 A structural block diagram of a time slot allocation system based on an adaptive weighted particle swarm optimization algorithm is provided for this application. DETAILED DESCRIPTION

[0078] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0079] In the first aspect, the present application provides a time slot allocation method based on an adaptive weighted particle swarm optimization algorithm, which takes a single network node in a data link network as a particle, and obtains an overall optimal time slot allocation scheme for the network node by continuously moving the particle toward the optimal solution, such as Figure 1 As shown, the following steps are included:

[0080] Step 1: Determine service parameters according to the service requirements of each network node in the data link network, wherein the service parameters include: the number of nodes with time slot requirements m, the total number of idle time slots S and the time slot resource requirement n;

[0081] Step 2: Initialize the algorithm parameters, including: particle swarm size M, particle position X, particle velocity V, learning factors C1 and C2, inertia weight range [W min ,W max ], the range of particle flight speed [v min ,v max ] and the maximum number of iterations K;

[0082] Step 3: Determine the interval between adjacent time slots of a single node Is the difference between the adjacent time slot interval P of the theoretical optimal time slot allocation less than or equal to the minimum value min? If so, jump to step 8, otherwise calculate the delay jitter when a single node allocates time slot resources;

[0083] The specific calculation process is as follows:

[0084] First, the time frame T with a period length of N f , denoted as T f ={T1,T2,...T i ...,T N}, where T i Indicates the position of the i-th time slot in a time frame. The idle time slot is L, L = {I1, I2, ..., I s}, (S≤N) means there are S idle time slot resources in N. When N=S, that is, T f is equivalent to L, indicating that a time frame with a period length of N will have N idle time slots. Now assume that n is randomly selected from S idle time slots. j The optimal time slot is allocated to node j, which can be expressed as: j ={x 1j ,x 2j ,...,x nj}(n≤S≤N), where X j It is called a time slot allocation scheme. In the time slot allocation scheme, the interval between adjacent time slots of the optimal time slot allocation should be:

[0085] P=S / n j (1)

[0086] Then for the time slot allocation scheme X j For example, the distance between adjacent time slots actually allocated to node j should be:

[0087]

[0088] In formula (2), represents the interval between the i+1th and ith time slots assigned to node j in the kth iteration, where represents the position of the i-th time slot assigned to node j at the k-th iteration. When i=n, ​​that is, is the interval between the last time slot allocated in this time frame and the first time slot allocated to the node in the next time frame, given by The resulting set can be written as like When , it means that the time slot allocation scheme has reached the optimal state, the allocated time slot resources are in a completely uniform state and the delay jitter is 0. However, in actual application, the uniformity of the allocated time slots cannot reach this level when allocating time slots to each node combat member according to their business time slot requirements. Therefore, the time slot allocation scheme can only be optimized. Infinitely approach the adjacent time slot interval P value of the theoretical optimal time slot allocation. In specific implementation, a minimum value can be set based on experiments or experience, so that the difference between the actual time slot interval of the particle and the adjacent time slot interval of the theoretical optimal time slot allocation is less than or equal to the minimum value. The delay jitter var of node j j It can be solved by formula (3):

[0089]

[0090] In formula (3), is the interval between the ith adjacent time slots assigned to node j at the kth iteration, n j is the time slot resource requirement of the jth node. According to formula (1), P is the total idle time slots S in a frame divided by the number of time slots n that node j needs to allocate j , that is, P is the adjacent time slot interval of the theoretically optimal time slot allocation.

[0091] Step 4: Calculate the average value of the delay jitter of each node when multiple nodes allocate time slot resources;

[0092] Assume that there are m network node members in the data link network who need to send or receive business messages, n j is the time slot resource requirement for sending and receiving the service message of the jth node, so the total time slot requirement Num of each network node in the tactical data link max The formula to be satisfied is as follows:

[0093]

[0094] When allocating time slot resources for multiple nodes that have business message sending or receiving requirements, the delay jitter of the data link network should be the average value of the delay jitter of each node, such as Figure 4 As shown, the average delay jitter at this time, that is, the fitness function fit, should be: fit = min{fit k ,k=1,2,...K}where

[0095] In formula (5), m is the number of nodes with time slot requirements in the data link network, var j The delay jitter generated when allocating time slots to the above node j.

[0096] Step 5: Update the optimal time slot p of the network node according to the delay jitter results calculated in steps 3 and 4 best and the global optimal time slot g best ;

[0097] For each particle, the delay jitter result calculated by formula (3) is compared with the result of the last delay jitter of the particle, and the particle position corresponding to the smaller delay jitter is taken as the current optimal position of the particle, and the optimal time slot p of the node is updated accordingly. best ;

[0098] Compare the average delay jitter result output by formula (5) with the average delay jitter value of the previous iteration, and take the smaller average delay jitter as the value p best As the global optimal position, and update the global optimal time slot g accordingly best .

[0099] Step 6: Update the inertia weight W according to the adaptive weight particle swarm optimization algorithm;

[0100] The updating formula for the inertia weight in step 6 is as follows:

[0101] W=W max -(W max -W min )×(k / K) 2 (6)

[0102] In formula (6), W is the inertia weight, W max , W min They respectively represent the maximum and minimum values ​​of the inertia weight W, where k represents the number of iterations of the current program and K represents the maximum number of iterations.

[0103] Step 7: Update the particle speed and position according to the inertia weight updated in step 6;

[0104]

[0105] Where:

[0106] k is the number of iterations;

[0107] W is the inertia weight updated in step 5;

[0108] p i represents the optimal position of a single particle, which is equivalent to p best ;P g represents the optimal position of the global particle, which is equivalent to g best ;

[0109] C1 and C2 are learning factors, where C1 is the individual learning factor of each particle, and C2 is the individual social learning factor of each particle. C1 and C2 are numbers between 1 and 4, usually C1=C2=2; r1 and r2 are random numbers between 0 and 1;

[0110] The particle flight speed is limited to the range of -30≤v≤30, and when v>v max When the particle velocity exceeds the maximum velocity threshold, v = v max On the contrary, when v<-v max When v=v min ;

[0111] The particle position X is processed out of bounds: when X>S, X=S, when X<0, X=0, S is the number of idle time slots.

[0112] Add an anti-collision mechanism between particles. The particle anti-collision mechanism is to prevent particles from being too concentrated in the same position during the search process, which may cause the algorithm to fall into a local optimum or converge prematurely. The particle anti-collision mechanism is usually used to enhance the diversity and global search capability of the particle swarm algorithm, so that the particle swarm maintains a certain degree of dispersion.

[0113] In each iteration, a repulsive force is added to prevent particles from getting too close to each other based on the distance between the particles. The repulsive force can be calculated by the following formula:

[0114]

[0115] Where F is the repulsive force, c is a constant, d is the distance between the two particles, and the direction of the repulsive force is away from the position of the other particle. When the particles get closer to each other, the repulsive force increases, forcing them apart.

[0116] In general, the value of c can range from 0.1 to 10 or 0.01 to 1, depending on the scale of the problem and the way the particle distance is measured. For larger search spaces or larger distances, larger c values ​​(such as 1 to 10) are usually required; while for smaller search spaces or denser particle distributions, smaller c values ​​(such as 0.01 to 1) may be more appropriate. In practical applications, the value of c can be adjusted dynamically. For example, it can be adjusted according to the current degree of aggregation or average distance of the particle swarm. If the particle swarm as a whole tends to converge or is too concentrated, increase c; if the particle swarm is too dispersed or divergent, reduce c. In this algorithm, c is taken as 0.01.

[0117] Step 8: Repeat steps 3 to 7, and increase the number of iterations by 1. When the number of iterations reaches the maximum number of iterations, stop the iteration, and use the optimal position of each particle as the optimal time slot allocation scheme for each network node and output it.

[0118] When the number of iterations does not reach the maximum number of iterations, repeat steps 3 to 7 and increase the number of iterations by 1; when the number of iterations reaches the maximum number of iterations, stop the iterations, combine the time slot optimal position vectors into the time slot optimal allocation plan, and output the optimal and uniform time slot solution.

[0119] Second, as Fig. 9 As shown, the present application provides a time slot allocation system based on an adaptive weighted particle swarm optimization algorithm. The system uses a single network node in a data link network as a particle, and continuously moves the particle toward the optimal solution direction to obtain an overall optimal time slot allocation scheme for the network node, which includes:

[0120] The network initialization module 810 is used to determine service parameters according to the service requirements of each network node in the data link network, wherein the service parameters include: the number of nodes with time slot requirements m, the total number of idle time slots S and the time slot resource requirement n;

[0121] The particle swarm initialization module 820 is used to initialize the algorithm parameters, including: particle swarm size M, particle position X, particle velocity V, learning factors C1 and C2, inertia weight range [W min ,W max ], the range of particle flight speed [v min ,v max ] and the maximum number of iterations K;

[0122] The first calculation module 830 is used to determine the interval between adjacent time slots of a single node. Is the difference between the adjacent time slot interval P of the theoretical optimal time slot allocation less than or equal to the minimum value? If so, add 1 to the number of iterations; otherwise, calculate the delay jitter when a single node allocates time slot resources;

[0123] The second calculation module 840 is used to calculate the average value of the delay jitter of each node when multiple nodes allocate time slot resources;

[0124] The first updating module 850 is used to update the optimal time slot p of the network node according to the average value of the delay jitter calculated by the first calculation module and the delay jitter calculated by the second calculation module. best and the global optimal time slot g best ;

[0125] A second updating module 860 is used to update the inertia weight W according to an adaptive weight particle swarm optimization algorithm;

[0126] A third updating module 870, used to update the particle speed and position according to the inertia weight updated by the second updating module;

[0127] The optimal time slot allocation module 880 is used to repeatedly execute steps 3 to 7, and increase the number of iterations by 1. When the number of iterations reaches the maximum number of iterations, the iteration is stopped, and the optimal position of each particle is used as the optimal time slot allocation scheme for each network node and output.

[0128] Furthermore, the first calculation module 830 is specifically configured to execute the following formula steps:

[0129]

[0130] In formula (3), is the interval between the ith adjacent time slots assigned to node j at the kth iteration, n j is the time slot resource requirement of the jth node. According to formula (1), P is the total idle time slots S in a frame divided by the number of time slots n that node j needs to allocate j , that is, P is the adjacent time slot interval of the theoretically optimal time slot allocation.

[0131] Furthermore, the second calculation module 840 is specifically configured to execute the following formula steps:

[0132] fit = min {fit k ,k=1,2,..K}

[0133] in

[0134] In formula (5), m is the number of nodes with time slot requirements in the data link network, var j The delay jitter generated when allocating time slots to the above node j.

[0135] Furthermore, the first updating module 850 is specifically configured to perform the following method steps:

[0136] For each particle, the delay jitter result calculated by formula (3) is compared with the result of the last delay jitter of the particle, and the particle position corresponding to the smaller delay jitter is taken as the current optimal position of the particle, and the optimal time slot p of the node is updated accordingly. best ;

[0137] Compare the average delay jitter result output by formula (5) with the average delay jitter value of the previous iteration, and take the smaller average delay jitter as the value p best As the global optimal position, and update the global optimal time slot g accordingly best .

[0138] Furthermore, the second updating module 860 is specifically configured to execute the following formula steps:

[0139] W=W max -(W max -W min )×(k / K) 2 (6)

[0140] In formula (6), W is the inertia weight, W max , W min They respectively represent the maximum and minimum values ​​of the inertia weight W, where k represents the number of iterations of the current program and K represents the maximum number of iterations.

[0141] The third updating module 870 is specifically used to execute the following formula steps:

[0142]

[0143] Where:

[0144] k is the number of iterations;

[0145] W is the inertia weight updated in step 5;

[0146] P i represents the optimal position of a single particle, which is equivalent to p best ;p g represents the optimal position of the global particle, which is equivalent to g best ;

[0147] C1 and C2 are learning factors, where C1 is the individual learning factor of each particle, and C2 is the individual social learning factor of each particle. C1 and C2 are numbers between 1 and 4, usually C1=C2=2; r1 and r2 are random numbers between 0 and 1.

[0148] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the time slot allocation method based on the adaptive weighted particle swarm optimization algorithm as described in the first aspect are implemented.

[0149] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the time slot allocation method based on the adaptive weighted particle swarm optimization algorithm as described in the first aspect are implemented.

[0150] The present invention adopts an adaptive inertia weight update strategy based on a particle swarm optimization algorithm (PSO), avoids the particle swarm algorithm from being easily trapped in a local optimal solution due to an excessively large or small inertia weight during iteration, accelerates the convergence speed at the end of iteration, improves algorithm performance, can quickly and accurately find the optimal time slot allocation scheme for a data link, can greatly reduce delay jitter, can quickly and accurately find the optimal scheme for time slot allocation; and can effectively complete the work of single-node and multi-node time slot resource allocation, meet the real-time transmission of tactical business messages in a data link, thereby greatly improving the combat effectiveness of a tactical data link and the stability of information transmission.

[0151] The application effect of the present invention is described in detail below in conjunction with simulation.

[0152] The simulation software used in this application is MATLAB R2021a, in which the command line window prints the generated simulation data. The global optimal fitness data in the simulation is the delay jitter index in the application of the present invention, and the global optimal particle position data is the optimal time slot allocation scheme of the present invention. The interval between the optimal particle positions is the interval between the allocated time slots. The interval between the two time slots determines the uniformity of the time slot resources allocated by the time slot allocation algorithm.

[0153] Simulation and verification of time slot allocation method:

[0154] a. Feasibility of the method

[0155] Assume that the time slot demand of a network node is 5 time slots, and allocate 5 time slot resources to the node in 1536 time slots. The simulation experiment parameters of single-node time slot allocation of AWPSO-SAA algorithm are set as follows: total number of idle time slots S = 1536, particle swarm size M = 50, m = 1, n = 5, maximum number of iterations K = 200, maximum inertia weight W max =1.2, minimum inertia weight W min =0.6. The global delay jitter simulated in this experiment is the delay jitter calculated by equation (5). Figure 2As shown in the figure, the AWPSO-SAA time slot allocation method is superior to the genetic algorithm and the time division multiple access time slot allocation method based on the manta ray algorithm when allocating time slot resources for a single node. When the AWPSO-SAA algorithm iterates to 45 rounds, the delay jitter converges to 0.16. And the AWPSO-SAA algorithm iterates 200 rounds and only takes 0.367s. This application is far better than the results obtained by the time slot allocation algorithm based on the binary tree time slot pool, and can meet the real-time requirements of the tactical data link.

[0156] At this time, the AWPSO-SAA time slot allocation algorithm is used to allocate 5 time slots to a node member in the data link network. The theoretical optimal interval obtained by equation (1) should be: P = 307.2. The optimal time slot allocation schemes obtained by simulation are X AP ={307, 614, 921, 1229, 1536}, the time slot intervals are all: P ={307, 307, 308, 307, 307}; the time slot intervals obtained by allocating time slots by the AWPSO-SAA algorithm are consistent with the theoretical optimal time slot intervals. The algorithm has a small number of iterations and consumes less time to find the optimal time slot allocation solution. The obtained time slot intervals are uniform and close to the theoretical optimal time slot intervals. Compared with the classic binary tree time slot allocation algorithm, the binary tree time slot allocation algorithm will cause a waste of time slot resources when allocating time slot resources that are not a power of 2. The algorithm in this article can accurately allocate the required time slot resources according to the specific business needs of the network nodes. Therefore, the data link time slot allocation method based on the adaptive weighted particle swarm optimization algorithm in this application has certain advantages in the allocation of data link time slot resources.

[0157] like Figure 3 As shown, the AWPSO-SAA time slot allocation method proposed in this application is used to allocate time slot resources to multiple nodes in the data link. Assume that there are currently four network nodes A, B, C, and D that have the need to receive and send business messages. In 512 idle time slots, time slot resources are randomly allocated to these four nodes according to their corresponding business needs. The multi-node time slot allocation simulation parameters are set to: idle time slot S = 512, particle swarm number M = 300, m = 4, and the business time slot resource requirements of the four nodes are: n = {2, 3, 6, 8, number of iterations K = 300, inertia weight W max =1.2, W min=0.6. The global delay jitter at this time should be the average delay jitter obtained by formula (5). When the AWPSO-SAA time slot resource method iterates to 190 rounds, the average delay jitter converges to 0.174. At this time, the optimal time slot allocation schemes allocated to the four nodes A, B, C, and D are: X1 = {1, 257}, X2 = {80, 251, 421}, X3 = {40, 125, 210, 295, 381, 466}, X4 = {63, 127, 191, 256, 320, 384, 448, 512}. The time slot intervals allocated to nodes A, B, C, and D are: P1 = {256, 256}, P2 = {171, 170, 171}, P3 = {85, 85, 85, 86, 85, 86}, P4 = {64, 64, 65, 64, 64, 64, 63}. From formula (1), the theoretically optimal time slot intervals of the four nodes should be: P1 = 256, P2 = 170.67, P3 = 85.33, P4 = 64. The adjacent time slot intervals allocated to each node obtained by experimental simulation are infinitely close to the theoretical optimal time slot interval. Therefore, the AWPSO-SAA time slot allocation method proposed in this application is used to solve the problem of multi-node time slot resource allocation. It is both efficient and very stable. The results of time slot allocation tend to be optimal and meet the technical indicators of uniformity of adjacent time slot intervals. Therefore, the fast optimized time slot allocation method proposed in the present invention is very suitable for a real electronic countermeasure environment. The time slot allocation optimization method can quickly and dynamically optimize network time slot resources for allocation according to the specific time slot requirements of the service, and can achieve high-quality and effective message transmission.

[0158] b. Time complexity of time slot allocation method

[0159] In order to explore the influence of the number of particle swarms in the AWPSO-SAA time slot allocation algorithm of the present invention on the optimal solution for time slot allocation, an experiment is first designed to study the relationship between the number of particle swarms and the efficiency of the algorithm. The specific experimental parameters are set as follows: idle time slots S = 512, particle swarm number M = 300, iteration number K = 200, inertia weight W max =1.2, W min =0.6. The time cost diagram is as follows Figure 5 As shown in Figure 1, the AWPSO-SAA time slot allocation method with four different particle swarm numbers was experimentally solved 50 times to obtain the average of the time overhead, eliminating the interference of accidental factors and ensuring the reliability of the experiment. Figure 5It can be seen that when the idle time slot S=512, the number of particle swarms M=50, and the time slot demand n is 5, the experimental results show that the AWPSO-SAA time slot allocation algorithm proposed by the present invention consumes only 0.449s through 200 iterations. The algorithm itself has found the optimal time slot allocation scheme when it iterates about 40 times, and the time overhead is less than 0.449s. The AWPSO-SAA time slot allocation method is superior to the results obtained by the time slot allocation algorithm based on the binary tree time slot pool and the time division multiple access time slot allocation method based on the manta ray algorithm, and can meet the real-time requirements of the tactical data link.

[0160] As the number of particle swarms increases, the time overhead will also increase accordingly. To a certain extent, too large an M value will have a certain impact on the convergence speed and convergence effect of the algorithm. Therefore, when using the AWPSO-SAA time slot allocation algorithm to allocate resources, the appropriate number of particle swarms should be selected according to the specific business message time slot requirements.

[0161] c. Algorithm stability

[0162] The stability of the time slot allocation method is a crucial factor for reallocating time slot resources according to the specific changes in the battlefield in the electronic countermeasure environment. This paper verifies the stability of the algorithm by using multiple experiments to observe the number of optimal time slot allocation algorithms. The solution formula for stability is shown in formula (9). The larger the Stability value, the more stable the time slot allocation algorithm.

[0163]

[0164] Formula (9), Q best It is the number of experiments with the same set of parameters to obtain the optimal solution, and Q is the total number of experiments with the same set of parameters.

[0165] In order to verify the stability of the AWPSO-SAA time slot allocation method, a verification experiment is designed. The specific experimental parameters are set as follows: particle swarm number M = 50, node time slot demand n = 5, iteration number K = 200, inertia weight W max =1.2, W min = 0.6, where the idle time slots are S = {128, 256, 512, 1563}. Each set of data will be repeated 50 times, and then the frequency of the time slot allocation algorithm finding the optimal solution will be counted, and the stability of the algorithm under each set of data will be calculated through (9). The effect of the stability of the time slot allocation algorithm is shown in Figure 6As shown in the figure, the AWPSO-SAA algorithm can maintain high stability during the experiment, and can also obtain suboptimal solutions without obtaining the optimal solution. Therefore, the AWPSO-SAA time slot allocation algorithm proposed in this paper is very suitable for real electronic countermeasure environments. The optimization algorithm stably calculates the optimal time slot allocation scheme, which can better complete the dynamic optimization of network time slot resource allocation.

[0166] d. The impact of particle swarm number on algorithm delay

[0167] In order to explore the influence of the number of particle swarms in the AWPSO-SAA time slot allocation method proposed in this invention on the optimal solution for time slot allocation, the following simulation experiment is designed to study the relationship between the number of particle swarms and the delay jitter of single-node time slot allocation. The simulation experiment parameters are set as follows: idle time slot S = 100, single-node service time slot demand n = 15, and number of iterations K = 100. The experimental simulation results are shown in Figure 2. Figure 7 As shown, when the number of iterations k = 30, the delay jitter decreases as the number of particle swarms increases. Experiments show that when the idle time slots are constant, the more particle swarms there are, the better the uniformity of the time slots allocated by the present invention will be. And when the number of particle swarms reaches 450, the delay jitter converges to 0.2222. That is, when the number of particle swarms is greater than or equal to 30 times the number of allocated time slots, the iterative convergence speed of the AWPOS-SAA time slot allocation method proposed in this application is faster. At this time, the time slot allocation method can achieve maximum performance, that is, it can find the optimal time slot allocation scheme faster and more effectively.

[0168] e. The impact of the number of idle time slots on the algorithm delay

[0169] In order to explore the influence of the number of idle time slots on the time slot allocation efficiency and delay jitter when the AWPSO-SAA time slot allocation method of the present invention allocates time slot resources according to the business time slot requirements of the combat units. A simulation experiment is designed to study the relationship between the number of idle time slots and delay jitter. The parameters of this simulation experiment are set as follows: the business time slot requirement of a single node n=15, the number of particle groups is 30 times the time slot requirement M=450 mentioned in the above "d. The influence of the number of particle groups on the algorithm delay", and the number of iterations K=100. The experimental simulation results are as follows Figure 8 As shown, when the number of iterations k = 30, the delay jitter decreases with the increase of the number of idle time slots and finally tends to be flat. The experimental simulation results show that when the service time slot demand of the node is certain, the more idle time slots there are, the better the uniformity of the time slots allocated by the time slot allocation method proposed in the present invention will be.

[0170] In summary, the data link time slot allocation method based on the adaptive weighted particle swarm optimization algorithm proposed in this application has efficient optimization capability, can quickly and accurately find the optimal solution for time slot allocation, and can effectively deal with problems such as time slot resource allocation. This method can greatly reduce delay jitter and quickly find the optimal time slot allocation strategy, thereby greatly improving the combat effectiveness of tactical data links and the stability of information transmission.

[0171] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A time slot allocation method based on adaptive weighted particle swarm optimization algorithm, characterized in that: A single network node in the data link network is regarded as a particle, and the particle moves continuously toward the optimal solution direction, thereby obtaining the overall optimal time slot allocation scheme of the network node, which includes the following steps: Step 1: Determine service parameters according to the service requirements of each network node in the data link network, wherein the service parameters include: the number of nodes with time slot requirements m, the total number of idle time slots S and the time slot resource requirement n; Step 2: Initialize the algorithm parameters, including: particle swarm size M, particle position X, particle velocity V, learning factors C1 and C2, inertia weight range [W min ,W max ], the range of particle flight speed [v min ,v max ] and the maximum number of iterations K; Step 3: Determine the interval between adjacent time slots of a single node Is the difference between the adjacent time slot interval P of the theoretical optimal time slot allocation less than or equal to the minimum value min? If so, jump to step 8, otherwise calculate the delay jitter when a single node allocates time slot resources; Step 4: Calculate the average value of the delay jitter of each node when multiple nodes allocate time slot resources; Step 5: Update the optimal time slot p of the network node according to the delay jitter results calculated in steps 3 and 4 best and the global optimal time slot g best ; Step 6: Update the inertia weight W according to the adaptive weight particle swarm optimization algorithm; Step 7: Update the particle speed and position according to the inertia weight updated in step 6; Step 8: Repeat steps 3 to 7, and increase the number of iterations by 1. When the number of iterations reaches the maximum number of iterations, stop the iteration, and use the optimal position of each particle as the optimal time slot allocation scheme for each network node and output it.

2. The time slot allocation method based on adaptive weighted particle swarm optimization algorithm according to claim 1, characterized in that: The specific method for calculating the delay jitter when a single node allocates time slot resources in step 3 is: is the interval between the ith adjacent time slots assigned to node j at the kth iteration, n j is the time slot resource requirement of the jth node. According to formula (1), P is the total idle time slots S in a frame divided by the number of time slots n that node j needs to allocate j , that is, P is the adjacent time slot interval of the theoretically optimal time slot allocation.

3. The time slot allocation method based on adaptive weighted particle swarm optimization algorithm as claimed in claim 2, characterized in that: The formula for calculating the average value of the delay jitter of each node when multiple nodes allocate time slot resources in step 4 is: fit = min {fit k ,k=1,2,…K} in The fit is a fitness function, which represents the average delay jitter of each node, m is the number of nodes with time slot requirements in the data link network, and var j The delay jitter generated when allocating time slots to the above node j.

4. The time slot allocation method based on adaptive weighted particle swarm optimization algorithm as claimed in claim 3, characterized in that: In step 5, the optimal time slot p of the node is obtained. best and the global optimal time slot g best The method is: For each particle, the delay jitter result calculated by formula (3) is compared with the result of the last delay jitter of the particle, and the particle position corresponding to the smaller delay jitter is taken as the current optimal position of the particle, and the optimal time slot p of the node is updated accordingly. best ; Compare the average delay jitter result output by formula (5) with the average delay jitter value of the previous iteration, and take the smaller average delay jitter as the value p best As the global optimal position, and update the global optimal time slot g accordingly best .

5. The time slot allocation method based on adaptive weighted particle swarm optimization algorithm as claimed in claim 4, characterized in that: The updating formula for the inertia weight in step 6 is as follows: W=W max -(W max -W min )×(k / K) 2 (6) In formula (6), W is the inertia weight, W max , W min They respectively represent the maximum and minimum values ​​of the inertia weight W, where k represents the number of iterations of the current program and K represents the maximum number of iterations.

6. The time slot allocation method based on adaptive weighted particle swarm optimization algorithm as claimed in claim 5, characterized in that: The velocity update formula and position update formula for the particle in step 7 are (7) and (8) respectively: Where: k is the number of iterations; W is the inertia weight updated in step 5; p i represents the optimal position of a single particle, which is equivalent to p best ;p g represents the optimal position of the global particle, which is equivalent to g best ; C1 and C2 are learning factors, where C1 is the individual learning factor of each particle and C2 is the individual social learning factor of each particle; r1 and r2 are random numbers between 0 and 1.

7. The time slot allocation system based on adaptive weighted particle swarm optimization algorithm is characterized by: The system takes a single network node in the data link network as a particle, and through the particles constantly moving towards the optimal solution, the overall optimal time slot allocation scheme of the network nodes is obtained, which includes: A network initialization module, used to determine service parameters according to the service requirements of each network node in the data link network, wherein the service parameters include: the number of nodes with time slot requirements m, the total number of idle time slots S and the time slot resource requirement n; The particle swarm initialization module is used to initialize the algorithm parameters, including: particle swarm size M, particle position X, particle velocity V, learning factors C1 and C2, inertia weight range [W min ,W max ], the range of particle flight speed [v min ,v max ] and the maximum number of iterations K; The first calculation module is used to calculate the delay jitter when a single node allocates time slot resources; The second calculation module is used to calculate the average value of the delay jitter of each node when multiple nodes allocate time slot resources; A first updating module is used to update the optimal time slot p of the network node according to the average value of the delay jitter calculated by the first calculation module and the delay jitter calculated by the second calculation module. best and the global optimal time slot g best ; The second updating module is used to update the inertia weight W according to the adaptive weight particle swarm optimization algorithm; A third updating module, used for updating the particle speed and position according to the inertia weight updated by the second updating module; The optimal time slot allocation module is used to repeatedly execute steps 3 to 7, and increase the number of iterations by 1. When the number of iterations reaches the maximum number of iterations, the iteration is stopped, and the optimal position of each particle is used as the optimal time slot allocation scheme of each network node and outputted.

8. The time slot allocation system based on adaptive weighted particle swarm optimization algorithm as claimed in claim 7, characterized in that: The second updating module is specifically used to update the inertia weight according to formula (6): W=W max -(W max -W min )×(k / K) 2 (6) In formula (6), W is the inertia weight, W max , W min They respectively represent the maximum and minimum values ​​of the inertia weight W, where k represents the number of iterations of the current program and K represents the maximum number of iterations.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the time slot allocation method based on the adaptive weighted particle swarm optimization algorithm as described in any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the time slot allocation method based on the adaptive weighted particle swarm optimization algorithm as described in any one of claims 1 to 6 are implemented.

Images