Edge air defense node task allocation method and system based on genetic algorithm and contract net method

Abstract

The invention discloses an edge air defense node task allocation method and system based on a genetic algorithm and a contract net method. The method specifically comprises the following steps: establishing a target threat degree estimation and task allocation model; the method comprises the following steps: calculating a multi-platform cooperative task allocation problem by adopting an improved mixed single parent genetic algorithm, through genetic operator design of integer coding + dynamic double populations, tournament selection + reverse-order crossover + adaptive variation and combining elite pool updating of a simulated annealing criterion and a mutation strategy triggered by concentration; aiming at the problem when an air defense platform encounters an emergency situation during task execution, an improved contract net method is adopted, and dynamic air defense task redistribution is realized by expanding a contract protocol process, introducing a multi-Agent collaborative architecture, designing a consistent auction algorithm and a dynamic bidding correction rule and combining a load-sensitive contract exchange mechanism. According to the method, multi-platform cooperative air defense task allocation can be rapidly and effectively carried out, the method dynamically adapts to environmental changes, and the combat effectiveness and combat efficiency of the multi-platform cooperative air defense tasks are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent optimization and distributed collaborative decision-making, and particularly to a method and system for task allocation of edge air defense nodes based on genetic algorithms and contract net methods. Background Art

[0002] The distributed air defense mission system integrates various types of data such as deduction results, daily training, assessment, and exercises, and uses operational calculation methods to quantitatively describe its own capabilities and mission requirements. Under the condition of information sharing in network-centric warfare, the system can quickly perceive targets and formulate efficient allocation strategies to achieve the best combat defense effect. Its main construction contents include algorithm models and software systems. The algorithm models cover key links such as self-state assessment, task allocation model construction, objective function design, and solution selection to ensure that the system can accurately track and intercept main targets.

[0003] Patent invention CN 118296479 A discloses an interception ballistic missile multi-target decision-making algorithm based on a covariance adaptive model. By based on a given offensive ballistic trajectory, predicted hit points, and threat degree evaluation index set, and calculating and sorting the threat degree values of ballistic missile targets based on the obtained data, and based on the decision variable allocation with constrained conditions and the covariance adaptive sampling strategy, parameters such as the launch azimuth angle, maximum negative attack angle, and launch position of the interceptor corresponding to different targets are determined, improving the interception efficiency and accuracy in multi-target missile interception. However, this algorithm does not have cross-platform collaborative air defense, fails to make full use of edge air defense nodes for multi-platform collaborative air defense, and cannot maximize the combat effectiveness of the system. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for task allocation of edge air defense nodes based on genetic algorithms and contract net methods, so as to quickly, effectively, and dynamically adapt to environmental changes for distributed air defense and anti-missile task allocation, and improve the flexibility of weapon units and the combat effectiveness and efficiency of edge air defense node task allocation methods.

[0005] The technical solution for achieving the purpose of the present invention is: A method for task allocation of edge air defense nodes based on genetic algorithms and contract net methods, including the following steps:

[0006] Step 1: Establish an estimated target threat degree and task allocation model;

[0007] Step 2: Adopt an improved hybrid single-parent genetic algorithm to calculate the multi-platform collaborative task allocation problem; The improved hybrid single-parent genetic algorithm designs genetic operators through integer coding + dynamic double populations, tournament selection + reverse order crossover + adaptive mutation, and combines the elite pool update based on the simulated annealing criterion and the mutation strategy triggered by concentration.

[0008] Step 3: For the problems encountered during the mission execution of the air defense platform in case of emergencies, the improved contract net method is used to solve the task reallocation problem; the improved contract net method realizes dynamic air defense task reallocation by expanding the contract protocol process, introducing a multi-Agent collaborative architecture, designing a consistency auction algorithm and a dynamic bid correction rule, and combining a load-sensitive contract exchange mechanism.

[0009] An edge air defense node task allocation system based on genetic algorithm and contract net method, which is used to implement the edge air defense node task allocation method based on genetic algorithm and contract net method. The system includes a first module to a third module, and the functions of each module are as follows:

[0010] The first module is used to establish a target threat degree estimation and task allocation model;

[0011] The second module uses an improved hybrid single-parent genetic algorithm to calculate the multi-platform collaborative task allocation problem; the improved hybrid single-parent genetic algorithm designs genetic operators through integer coding + dynamic dual populations, tournament selection + reverse order crossover + adaptive mutation, and combines an elite pool update based on simulated annealing criterion and a mutation strategy triggered by concentration;

[0012] The third module: For the problems encountered during the mission execution of the air defense platform in case of emergencies, the improved contract net method is used to solve the task reallocation problem; the improved contract net method realizes dynamic air defense task reallocation by expanding the contract protocol process, introducing a multi-Agent collaborative architecture, designing a consistency auction algorithm and a dynamic bid correction rule, and combining a load-sensitive contract exchange mechanism.

[0013] A mobile terminal includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the edge air defense node task allocation method based on genetic algorithm and contract net method.

[0014] Compared with the prior art, the significant advantages of the present invention are: (1) It can quickly and effectively perform multi-platform collaborative air defense task allocation, dynamically adapt to environmental changes, and the weapon units are more flexible and resilient; (2) An auction algorithm is used to implement task pre-planning. On the basis of the traditional contract net, a task transfer station and task priorities are introduced to assist in processing the task allocation scheme. By using exchange contracts, real-time collaboration allocation is carried out again on the basis of the original allocation scheme, improving the combat effectiveness and combat efficiency. Description of the Drawings

[0015] Figure 1 It is a flowchart of the edge air defense node task allocation method based on genetic algorithm and contract net method of the present invention.

[0016] Figure 2 This is a schematic flow chart of the improved hybrid parthenogenetic algorithm in the present invention.

[0017] Figure 3 This is a schematic flow chart of the improved contract net algorithm in the present invention. Specific embodiments

[0018] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] Combined with Figure 1 , a method for task allocation of edge air defense nodes based on genetic algorithm and contract net method in the present invention includes the following steps:

[0020] Step 1: Establish a target threat degree estimation and task allocation model;

[0021] Step 2: Use the improved hybrid parthenogenetic algorithm to calculate the multi-platform collaborative task allocation problem; the hybrid parthenogenetic algorithm designs genetic operators through integer coding + dynamic dual populations (ordinary population and elite pool), tournament selection + reverse order crossover + adaptive mutation, and combines the elite pool update based on simulated annealing criterion and the mutation strategy triggered by concentration degree, significantly improving the solution efficiency and global convergence of the task allocation problem, and is especially suitable for the collaborative optimization scenario of air defense nodes with multiple constraints and high discreteness.

[0022] Step 3: In response to the problem of sudden situations encountered during the execution of tasks by air defense platforms, use the improved contract net method to solve the task reallocation problem. By expanding the contract protocol process, introducing a multi-Agent collaborative architecture (air defense / task / management Agent), designing a consistency auction algorithm and a dynamic bidding correction rule, and combining a load-sensitive contract exchange mechanism, the real-time performance, load balancing and global collaborative efficiency of the contract net method in dynamic air defense task reallocation are significantly improved.

[0023] As a specific example, the edge air defense node task is as follows: a certain military defense key area faces N incoming missiles, and there are M of our air defense weapon units around to intercept. The following constraints need to be considered when allocating targets:

[0024] All incoming missiles must be assigned weapons for strike;

[0025] Each missile can only be intercepted by one weapon platform;

[0026] The number of tasks of the air defense platform cannot exceed its capacity;

[0027] The air defense platform can only intercept one missile at the same time;

[0028] When the number of platforms is less than the number of missiles, give priority to intercepting missiles with a higher threat level.

[0029] As a specific example, the establishment of the target threat degree estimation and task assignment model described in step 1 is as follows:

[0030] Step 1.1: Define the attribute sets of the air defense Agent and missile information as:

[0031]

[0032] In the formula, represents the attribute set of the i-th air defense weapon unit, represents the attribute set of the j-th target missile, represents the coordinates (x i , y i , z i ) of the i-th air defense weapon unit, represents the launch speed of the i-th air defense weapon unit, represents the integrity of the i-th air defense weapon unit, represents the launch speed of the i-th air defense weapon unit, represents the coordinates (x j , y j , z j ) of the j-th target missile;

[0033] respectively represent the value, speed, course angle, and coverage radius of the j-th target missile;

[0034] Step 1.2: Establish the objective function model for agent collaborative task assignment as:

[0035] Max[f1(X)-f2(X)](3)

[0036] In the formula, f1(X) is the benefit obtained by the agent for completing the task; f2(X) is the cost paid by the agent for completing the task, and the specific definitions are as follows:

[0037]

[0038] such that

[0039]

[0040] Equation (5) represents the uniqueness constraint, that is, a missile can be assigned at most once; the decision variable matrix X = {x ij |i = 1, 2…n; j = 1, 2…k}, where x ij is 1 indicating that the target missile j is assigned to the i-th air defense unit, and 0 indicating non-assignment; Reward(WA i (T j )) represents the benefit function of the i-th air defense platform destroying the j-th target missile, Cost(WAi (T j )) represents the i-th air defense platform WA i Destroy the jth target missile T j The cost estimation function of

[0041] Step 1.3, attack benefit refers to the mission value that the air defense platform WA can obtain when performing a certain mission. The attack benefit function is designed as:

[0042]

[0043] In the formula, WA i refers to the i-th air defense platform, T j represents the jth target missile, Reward(wA i (T j )) represents the profit function of the i-th air defense platform destroying the j-th target missile, α is the weight coefficient of the profit function, and α∈[0,1] represents the proportion of the profit function to the target; The missile's own value, usually determined by its model and destructive capability; P ij means WA i The probability of successfully destroying missile j;

[0044] In order to simplify the model, the damage probability P ij It can be obtained by mapping the polar coordinate equation of the current state of the air defense platform and the missile:

[0045]

[0046] Where ρ(i,j) represents the current air defense platform WA i The relative distance to the incoming missile j, It represents the derivative of relative distance with respect to time. When WA is launched and destroys the target, ρ = 0; q is the sight angle; are the launch speed of the air defense weapon unit and the speed of the target missile respectively; θ i ,θ j WA i The angle between the velocity vector of the j-th target missile and the horizontal reference line in the same vertical plane;

[0047] According to the quasi-parallel approach rule, when The larger it is, the higher the probability that the agent will intercept the target.

[0048] Step 1.4: Design the cost estimation function as:

[0049] Cost(WA i (T j ))=β×ρ(i,j)×x ij(8)

[0050] In the formula, Cost(WA i (T j )) represents the cost estimation function for the i-th air defense platform WA i to destroy the j-th target missile T j . x ij is 1 if mission j is assigned to air defense unit i, and 0 if not; ρ(i, j) represents the relative distance between the current air defense platform WA i and the incoming missile j; β represents the weight coefficient, β ∈ [0, 1];

[0051] Step 1.5: Design a target threat degree estimation model, which is specifically as follows:

[0052] Range threat factor μ L is:

[0053]

[0054] In the formula, L is the estimated range of the incoming missile;

[0055] Flight speed threat factor μ v is:

[0056] μ v = 1 - e -γv (10)

[0057] In the formula, γ = 0.0015, and v represents the missile flight speed;

[0058] Take the current distance between the incoming missile and our defense key area as an influencing factor for target threat estimation. The missile-target distance threat factor μ r is:

[0059]

[0060] In the formula, r is the missile-target distance. When r approaches the safety radius R of the defense key area from ∞, the choice of the constant k determines the rate of function decline. When k is larger, the rate of function decline is faster, and the closer the missile is to the key area, the more obvious the increase in threat degree;

[0061] Sum the range threat factor μ L , the flight speed threat factor μ v , and the missile-target distance threat factor μ r with weights to obtain the estimation function for the threat degree of the target missile T j , as shown below:

[0062] Threat j = ω L μ L + ωv μ v + ω r μ r , ω L + ω v + ω r = 1 (12)

[0063] In the formula, Threat j represents the threat level of the j-th target missile, ω L , ω v , ω r are the weight coefficients of the range threat factor μ L , the flight speed threat factor μ v , and the missile-target distance threat factor μ r . By adjusting the proportion of ω L , ω v , ω r , it is beneficial for the air defense Agent to judge the priority of the interception task according to the actual environment.

[0064] As a specific example, the improved hybrid single-parent genetic algorithm described in step 2 is used to calculate the multi-platform collaborative task allocation problem, as Figure 2 shown below:

[0065] Step 2.1: Adopt an integer coding scheme to map the integer decision vector into a chromosome, map the decision variable into a gene position, and each chromosome corresponds to a solution of the model;

[0066] Step 2.2: Population initialization:

[0067] Ordinary population: Randomly generated before the start of evolution;

[0068] Elite pool: The optimal individual or its mutant of each generation of the ordinary population;

[0069] Population size: Dynamically increasing in the early stage and reaching a fixed size in the later stage, usually between dozens and hundreds;

[0070] Step 2.3: Calculate the fitness function: The fitness function consists of the optimization objective function expression benefit and the cost function of multi-agent collaborative target allocation, as shown in the following formula:

[0071]

[0072] Where T j represents the gene set of the j-th chromosome segment; U i (T j ) represents that the weapon platform i executes the task T j .

[0073] Step 2.3: Design genetic operators, including a selection strategy operator, an inverse order crossover operator, and a transposition mutation operator, and perform genetic operations as follows:

[0074] Step 2.3.1: Selection strategy operator: Arrange the population in descending order according to fitness, select the optimal individual among them, replace the individual with the worst fitness in the population, and at the same time select the next-generation population through the competition of multiple contestants;

[0075] Specific operations:

[0076] (1) Determine the size of the tournament: Select an appropriate tournament size, that is, the number of individuals participating in the competition each time;

[0077] (2) Calculate the fitness f of each individual in the population. M is the population size;

[0078] (3) Conduct multiple tournament selections: Repeat the following steps until the new population is full:

[0079] a. Randomly select a group of individuals as tournament participants, and the number of individuals is equal to the tournament size;

[0080] b. Evaluate the fitness among the tournament participants and select the individual with the highest fitness as the winner of the tournament;

[0081] c. Add the winner of the tournament to the new population;

[0082] (4) Sort all individuals, select the individual with the highest fitness for replication, and replace the individual with the lowest fitness in the population;

[0083] Step 2.3.2: Inverse order crossover operator: Select a parent chromosome, then randomly select two different random numbers less than the number of target missiles, and reverse the genes between the gene positions represented by the two random numbers;

[0084] Calculation formula for inverse order crossover:

[0085]

[0086] Among them, σ is the differential coefficient of e σx Also known as the attenuation parameter, with a value range of 0 to 1, reflecting the trend of change in P c probability; P c0 is the initial crossover probability; Count is the number of consecutive occurrences of the current optimal fitness value; generation is the current generation number; iter_Max is the maximum number of iterations;

[0087] Step 2.3.3, Transposition Mutation Operator: To prevent individuals from falling into local optima, a mutation operator is introduced. An adaptive transposition mutation operator is adopted to enhance the population's ability to jump out of local optima;

[0088] Set the mutation probability of the transposition mutation operator as P m , and the calculation formula is:

[0089]

[0090] where τ is the differential coefficient of e -τx , with a value range of 0 to 5; P m0 is the initial mutation probability; Count is the number of consecutive occurrences of the current optimal fitness value; generation is the current generation number; iter_Max is the maximum number of iterations;

[0091] The mutation operation is carried out by randomly selecting two adjacent gene positions and swapping their gene values;

[0092] Step 2.4, Check whether the termination condition is met. If so, output the allocation result; otherwise, update the population and return to Step 2.2 for the next generation of genetic operations.

[0093] As a specific example, the design of the elite pool described in Step 2.2 is as follows:

[0094] Enhance the local optimization ability of the algorithm. It is composed of the ordinary population and its mutants in each generation. When the capacity Maxsize of the elite pool has not reached the upper limit, the number of individuals in the elite pool increases dynamically; when the number of individuals reaches the capacity limit, the elite pool starts to continuously detect the fitness of the individuals in the pool; the elite pool has two strategies:

[0095] (1) Elite pool update: Continuously update according to the best individual transmitted by the ordinary population in each generation according to the Metropolis criterion in the simulated annealing algorithm;

[0096] (2) Elite pool mutation: When the concentration of individuals in the population reaches a certain threshold, that is, the diversity decreases to a certain extent, the elite pool mutation is triggered.

[0097] As a specific example, the elite pool update strategy is as follows:

[0098] When the elite pool reaches the capacity limit, introduce the Metropolis criterion in the simulated annealing algorithm for update, and randomly search for the global optimal solution of the objective function in the solution space by combining the probability jump characteristic, which is specifically manifested as:

[0099] During evolution, the ordinary population outputs the optimal individual of each generation, whose fitness is E(n). The elite pool copies this individual and performs forced transposition mutation. After mutation, the fitness of the new individual is E(n + 1), and the update operation is completed according to the probability P;

[0100] The calculation formula of the probability P is:

[0101]

[0102] In the formula, T is the current temperature;

[0103] If E(n + 1)>E(n), it indicates that the fitness has been improved. Replace the worst individual in the pool with the new individual and feedback it to the ordinary population. The ordinary population completes the secondary update according to its own selection strategy.

[0104] As a specific example, the mutation strategy of the elite pool is as follows:

[0105] When the elite pool reaches the capacity limit, detect the average fitness f of the population in the elite pool avg and the proximity between the individual with the maximum fitness f max ;

[0106] The triggering condition satisfies the following inequality:

[0107]

[0108] In the formula, γ is the density factor, [γ lb ,γ ub are its upper and lower limits, representing the concentration degree of individuals; generation is the current generation number, and iter_Max is the maximum number of iterations.

[0109] As a specific example, for the problem encountered when the air defense platform is performing tasks and sudden situations occur in step 3, the improved contract net method is used to solve the task reallocation problem, as Figure 3 shown below:

[0110] Step 3.1: Introduce the market contract mechanism into the online task allocation problem of the air defense platform and the intercept missile. The contract net protocol is extended to "tendering - bidding - winning - signing - status update". Its input parameters are the platform, missile information and the matching relationship of the task transmitted by the task pre - allocation, which are defined in the form of a triple as <WA ch ,Missile ch ,X>, where WA ch is the platform transmitted by the task pre - allocation, Missile ch is the missile information, and X is the matching relationship of the task;

[0111] Step 3.2: Define three types of Agents, namely air defense Agent, task Agent, and management Agent. All Agents operate and cooperate based on their own rules to complete task allocation, represented by the triple W A , T A , C A . Among them, the set W A of air defense Agents is In the formula, represents the Agent of the air defense weapon unit WA i , where i is the serial number of the air defense weapon unit, used to manage the resource attributes and bidding calculation of WA i ; each air defense Agent exists on its corresponding weapon unit; based on the state update mechanism in the extended contract net, W i A will, when encountering the following situations, transfer the information of WA i to the management Agent:

[0112] (1) WA i is out of contact or damaged;

[0113] (2) The task set of WA i has changed;

[0114] The set T A of task Agents is In the formula, represents the task of intercepting the Missile j missile, where j is the missile serial number. It is responsible for auctioning this task, generating a tender contract according to the missile attribute set, and transferring it to the management Agent. Once the task is successfully executed, the resources occupied by it will be immediately destroyed and recycled to avoid waste of resources;

[0115] The set C A of management Agents is used to manage and receive information including air defense Agents and task Agents. The contract trading area acts as the role of the management Agent, matching the tender information transmitted by the task Agent and the bidding information transmitted by the air defense Agent for resources;

[0116] Step 3.3: After generating the initial task set of each air defense weapon unit through the improved hybrid single-parent genetic algorithm, the air defense Agent and the task Agent enter the monitoring state to detect whether there are new missiles and damage to air defense weapon units. If so, enter Step 3.4;

[0117] Step 3.4: Feed back the newly emerged missile's own attributes and the air defense tasks of the damaged WA that cannot continue to complete the tasks to the management Agent as the tasks to be tendered, and incorporate them into the contract trading area;

[0118] Step 3.5: The contract trading area releases tender information according to the priority of the tasks;

[0119] Step 3.6: The air defense Agents deployed on each air defense unit calculate their own task benefits, evaluate the costs of executing tasks, release bidding information, and conduct task auctions in the contract trading area;

[0120] Step 3.7: The trading area announces the list of winners, and the winners update their own task sequences; if there are tasks that are not won, jump to Step 3.5, otherwise enter Step 3.8;

[0121] Step 3.8: Conduct local contract adjustments among the high-load air defense Agents to improve the global coordination benefits of multiple weapon units and complete the status update;

[0122] Step 3.9: The auction contract adopts the consistent auction algorithm. When the number of tasks to be tendered T stored in the contract trading area is greater than the number of currently bid-able air defense Agents, the tasks are auctioned in T batches, and the quantity of each batch is equal to the number of buyers, so that each buyer can successfully bid for an auction item; after the auction ends, update the task number T in the next round;

[0123] Step 3.10: When T is less than the number of currently bid-able air defense Agents, determine the winner according to the bid values of each Agent. When the number of tasks undertaken by WA i has reached the average load the original bid value F of this buyer for the next auction item j is corrected;

[0124] The calculation and bid value correction formula are as follows:

[0125]

[0126] where N is the total number of air defense weapon units, load i is the load of the WA i k is the coefficient for correcting the bid value F, and F ij is the corrected bid value of WA i for the auction item j;

[0127] Step 3.11: After each auction contract, the WA i with the highest task load and the WA j with the second highest task load of the same type as it conduct sharing and communication. If WA j shares a certain task in its own task sequence with WAi If the overall objective function value of the two can be increased after the exchange, the contract exchange mechanism is triggered, and the contract exchange that maximally increases the effectiveness is selected for signing.

[0128] The present invention also provides an edge air defense node task allocation system based on a genetic algorithm and a contract net method. The system is used to implement the edge air defense node task allocation method based on the genetic algorithm and the contract net method. The system includes a first module to a third module, and the functions of each module are as follows:

[0129] The first module is used to establish a target threat degree estimation and task allocation model;

[0130] The second module uses an improved hybrid single-parent genetic algorithm to calculate the multi-platform collaborative task allocation problem; the improved hybrid single-parent genetic algorithm designs genetic operators through integer coding + dynamic dual populations, tournament selection + reverse order crossover + adaptive mutation, and combines elite pool update based on the simulated annealing criterion and a mutation strategy triggered by concentration;

[0131] The third module, aiming at the problem when an unexpected situation occurs during the task execution of the air defense platform, uses an improved contract net method to solve the task reallocation problem; the improved contract net method expands the contract protocol process, introduces a multi-Agent collaborative architecture, designs a consistency auction algorithm and a dynamic bid correction rule, and combines a load-sensitive contract exchange mechanism to achieve dynamic air defense task reallocation.

[0132] The present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the edge air defense node task allocation method based on the genetic algorithm and the contract net method.

[0133] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art in this technical field, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A method for allocating tasks of edge air defense nodes based on genetic algorithm and contract network method, characterized in that: The following steps are involved: Step 1: Establish a target threat estimation and task allocation model; Step 2: Using an improved hybrid parthenogenetic algorithm to calculate the multi-platform collaborative task allocation problem; the improved hybrid parthenogenetic algorithm is designed by integer coding + dynamic dual population, tournament selection + reverse crossover + adaptive mutation genetic operators, combined with the elite pool update of simulated annealing criteria and the mutation strategy triggered by concentration; Step 3: To solve the problem of task reallocation when the air defense platform encounters unexpected situations during mission execution, an improved contract network method is used to solve the task reallocation problem; the improved contract network method expands the contract agreement process, introduces a multi-agent collaborative architecture, designs a consistent auction algorithm and dynamic bidding correction rules, and combines a load-sensitive contract exchange mechanism to achieve dynamic air defense task reallocation.

2. The edge air defense node task allocation method based on genetic algorithm and contract network method according to claim 1 is characterized in that: The edge air defense node tasks are specifically: A key military defense location faces N incoming missiles and is surrounded by M air defense weapon units that need to be intercepted. The following constraints must be considered when allocating targets: All incoming missiles must be countered by weapons assigned; Each missile can be intercepted by only one weapon platform; The number of air defense platform missions cannot exceed the platform's own capabilities; The air defense platform can only intercept one missile at a time; When the number of platforms is less than the number of missiles, priority is given to intercepting missiles with higher threat levels.

3. The edge air defense node task allocation method based on genetic algorithm and contract network method according to claim 1 is characterized in that: In step 1, a target threat estimation and task allocation model is established, as follows: Step 1.1, define the attribute set of air defense agent and missile information as follows: In the formula, Agent represents the intelligent agent, represents the attribute set of the i-th air defense weapon unit, represents the attribute set of the i-th target missile, represents the coordinates of the ith air defense weapon unit on the x, y, and z axes, represents the firing speed of the i-th air defense weapon unit, represents the integrity of the i-th air defense weapon unit, represents the firing speed of the i-th air defense weapon unit, represents the x, y, and z-axis components of the j-th target missile's coordinates; They represent the value, speed, heading angle, and coverage radius of the j-th target missile respectively; Step 1.2: Establish the objective function model of agent collaborative task allocation: Max[f1(X)-f2(X)](3) In the formula, f1(X) is the benefit obtained by the agent after completing the task; f2(X) is the cost paid by the agent after completing the task. The specific definitions are as follows: Make Formula (5) represents the uniqueness constraint, that is, a missile is assigned at most once; the decision variable matrix X = {x ij |i=1,2…n;j=1,2…k}, where x ij 1 indicates that target missile j is assigned to air defense unit i, and 0 indicates that it is not assigned; Reward(WA i (T j )) represents the revenue function of the i-th air defense platform destroying the j-th target missile, Cost(WA i (T j )) represents the i-th air defense platform WA i Destroy the jth target missile T j The cost estimation function of Step 1.3, attack benefit refers to the mission value that the air defense platform WA can obtain when performing a certain mission. The attack benefit function is designed as: In the formula, WA i refers to the i-th air defense platform, T j represents the jth target missile, Reward(WA i (T j )) represents the profit function of the i-th air defense platform destroying the j-th target missile, α is the weight coefficient of the profit function, and α∈[0,1] represents the proportion of the profit function to the target; The missile's own value, usually determined by its model and destructive capability; P ij means WA i The probability of successfully destroying missile j; In order to simplify the model, the damage probability P ij The polar coordinate equation mapping of the current state of the air defense platform and the missile is obtained: Where ρ(i,j) represents the current air defense platform WA i The relative distance to the incoming missile j, It represents the derivative of relative distance with respect to time. When the air defense platform WA launches and destroys the target, ρ = 0; q is the sight angle; are the launch speed of the air defense weapon unit and the speed of the target missile respectively; θ i ,θ j WA i The angle between the velocity vector of the j-th target missile and the horizontal reference line in the same vertical plane; According to the quasi-parallel approach rule, when The larger it is, the higher the probability that the agent will intercept the target. Step 1.4: Design the cost estimation function as: Cost(WA i (T j ))=β×ρ(i,j)×x ij (8) In the formula, Cost(WA i (T j )) represents the i-th air defense platform WA i Destroy the jth target missile T j The cost estimation function, x ij 1 means that task j is assigned to air defense unit i, and 0 means it is not assigned; ρ(i,j) represents the current air defense platform WA i The relative distance to the incoming missile j; β represents the weight coefficient, β∈[0,1]; Step 1.5: Design a target threat estimation model, as follows: Range Threat Factor μ L for: Where L is the estimated range of the incoming missile; Flight speed threat factor μ v for: μ v =1-e -γv (10) In the formula, γ = 0.0015, v represents the missile flight speed; The current distance between the incoming missile and our defense stronghold is used as the influencing factor of target threat estimation, and the missile-target distance threat factor μ r for: In the formula, r is the distance between the missile and the target. When r approaches the safety radius R of the defense key point from ∞, the choice of constant k determines the rate of function descent. The larger k is, the faster the function descent rate is. The closer the missile is to the key point, the more obvious the threat level increases. The range threat factor μ L , Flight speed threat factor μ v 、Projectile-target distance threat factor μ r Perform weighted summation to obtain the target missile T j The estimation function of threat level is as follows: Threat j =ω L m L +oh v m v +oh r m r Oh, oh L +oh v +oh r =1 (12) Where Threat j represents the threat level of the j-th target missile, ω L ,ω v ,ω r is the range threat factor μ L , Flight speed threat factor μ v 、Projectile-target distance threat factor μ r The weight coefficient is adjusted by L ,ω v ,ω r The proportion will help the air defense agent to judge the priority of the interception task according to the actual environment.

4. The edge air defense node task allocation method based on genetic algorithm and contract network method according to claim 1 is characterized in that: In step 2, an improved hybrid parthenogenetic algorithm is used to calculate the multi-platform collaborative task allocation problem, as follows: Step 2.1, using an integer encoding scheme, map the integer decision vector into chromosomes, and the decision variables into gene positions, with each chromosome corresponding to a solution of the model; Step 2.2, population initialization: Normal population: randomly generated before evolution begins; Elite pool: the best individuals or their variants in each generation of the normal population; Population size: Dynamic growth in the early stage and a fixed size in the later stage; Step 2.3, calculate the fitness function: The fitness function is composed of the optimization objective function expression benefit and cost function of multi-agent collaborative target allocation, as shown in the following formula: Where T j represents the gene set of the jth chromosome segment; U i (T j ) indicates that weapon platform i performs mission T j ; Step 2.4: Design genetic operators, including selection strategy operator, reverse crossover operator and transposition mutation operator, to perform genetic operations, as follows: Step 2.4.1, select the strategy operator: sort the population in descending order according to the fitness, select the best individual, replace the individual with the worst fitness in the population, and select the next generation of population based on the competition among multiple contestants, as follows: (1) Determine the size of the tournament: Choose the tournament size, that is, the number of individuals participating in each selection; (2) Calculate the fitness f of each individual in the population, where M is the population size; (3) Perform multiple tournament selections: Repeat steps a to c until the new population is full: a. Randomly select a group of individuals as participants in the tournament, the number of individuals is equal to the size of the tournament; b. Evaluate the fitness among the tournament participants and select the individual with the highest fitness as the winner of the tournament; c. Add the winner of the tournament to the new population; (4) Sort all individuals and select the individuals with the highest fitness to replicate, replacing the individuals with the lowest fitness in the group; Step 2.4.2, reverse crossover operator: select a parent chromosome, then randomly select two different random numbers that are less than the target number of missiles, and reverse the order of the genes between the gene positions represented by the two random numbers; The calculation formula for reverse crossover is: Where σ is e σx The differential coefficient, also called the attenuation parameter, ranges from 0 to 1 and reflects P c Probability change trend; P c0 is the initial crossover probability; Count is the number of consecutive occurrences of the current optimal fitness value; generation is the current generation; iter_Max is the maximum number of iterations; Step 2.4.3, transposition mutation operator: In order to prevent individuals from falling into the local optimum, a mutation operator is introduced, and an adaptive transposition mutation operator is used to enhance the ability of the population to jump out of the local optimum; Set the mutation probability of the transposition mutation operator to P m , the calculation formula is: Where τ is e -τx The differential coefficient of P ranges from 0 to 5; m0 is the initial mutation probability; The mutation operation uses the method of randomly selecting two adjacent gene positions and exchanging gene values ​​to perform the mutation operation; Step 2.5: Check whether the termination condition is met. If so, output the allocation result; otherwise, update the population and return to step 2.2 for the next generation of inheritance.

5. The edge air defense node task allocation method based on genetic algorithm and contract network method according to claim 4 is characterized in that: The design of the elite pool in step 2.2 is as follows: Enhance the local optimization ability of the algorithm. The common population of each generation and its variants are composed of the elite pool. When the capacity Maxsize of the elite pool has not reached the upper limit, the number of individuals in the elite pool is dynamically increased; when the number of individuals reaches the upper limit of the capacity, the elite pool begins to continuously detect the fitness of the individuals in the pool; The elite pool has two strategies: (1) Elite pool update: Continuously update the best individuals of each generation from the general population according to the Metropolis criterion in the simulated annealing algorithm; (2) Elite pool mutation: When the individual concentration of a population reaches a threshold, that is, the diversity is reduced to a set level, triggering an elite pool mutation.

6. The edge air defense node task allocation method based on genetic algorithm and contract network method according to claim 5 is characterized in that: The elite pool update strategy is as follows: When the elite pool reaches its capacity limit, the Metropolis criterion in the simulated annealing algorithm is introduced for updating. By combining the probability jump characteristics, the global optimal solution of the objective function is randomly found in the solution space. The specific performance is as follows: During the evolution period, the common population outputs the best individual of each generation, whose fitness is E(n). The elite pool copies the individual and forces it to undergo transposition mutation. After the mutation, the fitness of the new individual is E(n+1), and the update operation is completed according to probability P. The calculation formula of probability P is: Where T is the current temperature; If E(n+1)>E(n), it means that the fitness has been improved. The worst individual in the pool is replaced by a new individual and fed back to the normal population. The normal population completes the secondary update according to its own selection strategy.

7. The edge air defense node task allocation method based on genetic algorithm and contract network method according to claim 5 is characterized in that: The elite pool mutation strategy is as follows: When the elite pool reaches its capacity limit, the average fitness f of the population in the elite pool is detected avg With the maximum fitness individual f max The degree of proximity between The trigger condition satisfies the following inequality: Where γ is the density factor, [γ lb ,γ ub ] are the upper and lower limits of γ, representing the degree of individual concentration; generation is the current generation, and iter_Max is the maximum number of iterations.

8. The edge air defense node task allocation method based on genetic algorithm and contract network method according to claim 1 is characterized in that: In step 3, in order to solve the problem of emergency when the air defense platform is performing tasks, the improved contract network method is used to solve the task reallocation problem, as follows: Step 3.1: Introduce the market contract mechanism into the online task allocation problem of air defense platforms and interceptor missiles. The contract network protocol is extended to: bidding-bid-winning-signing-status update. The input parameters are the platform and missile information transmitted by the task pre-allocation and the matching relationship of the task, which are defined in the form of a triple: <WA ch ,Missile ch ,X>, where WA ch Pre-allocate the platform for task delivery, Missile ch is the missile information, X is the matching relationship of the task; Step 3.2, define three types of agents: air defense agent, task agent, and management agent. All agents run and collaborate based on their own rules to complete task allocation. Use the triple W A , T A , C A To represent, where the set of air defense agents W A for In the formula, Indicates air defense weapon unit WA i Agent, i is the air defense weapon unit number, used to manage WA i Resource attributes and bidding calculations; each air defense agent exists on the corresponding weapon unit; Based on the state update mechanism in the extended contract network, WA will be used when the following situations occur i The information is passed to the management agent: (1)WA i Lost or destroyed; (2)WA i The mission set has changed; The set of task agents T A for In the formula, Indicates interception of Missile j The task of the missile is responsible for auctioning the task, generating a bidding contract based on the missile attribute set, and passing it to the management agent. Once the task is successfully executed, The occupied resources are immediately destroyed and recycled, j is the missile serial number; Management Agent Collection C A Used to manage and receive information including air defense agent and task agent. The contract transaction area acts as a management agent to match the bidding information from the task agent and the bidding information from the air defense agent. Step 3.3: After generating the initial task set of each air defense weapon unit, the air defense agent and the task agent enter the monitoring state to detect whether there are new missiles and air defense weapon units damaged. If so, proceed to step 3.4; Step 3.4: Feedback the attributes of the newly emerged missiles and the air defense mission of the damaged WA that cannot continue to complete the mission as tasks to be bid to the management agent and include them in the contract transaction area; Step 3.5: The contract transaction area publishes bidding information according to the priority of the task; Step 3.6: The air defense agent deployed on each air defense unit calculates its own mission benefits, evaluates the cost of executing the mission, publishes bidding information, and conducts mission auctions in the contract trading area; Step 3.7: The trading area announces the list of successful bidders, and the successful bidders complete the update of their own task sequences; if there are unsuccessful tasks, jump to step 3.5, otherwise go to step 3.8; Step 3.8: Local contract adjustments are made between high-load air defense agents to improve the global synergy benefits of multiple weapon units and complete the status update; Step 3.9, the auction contract uses a consistent auction algorithm. When the number of tasks to be bid T stored in the contract transaction area is greater than the number of air defense agents that can be bid, the number of tasks will be auctioned in T batches, and the number of each batch is equal to the number of buyers, so that each buyer can successfully bid for an auction item; the auction ends and the number of tasks T is updated in the next round; Step 3.10: When T is less than the number of air defense agents that can bid, the winner is determined based on the bid value of each agent. i The number of tasks undertaken has reached the average load When , the buyer modifies the original bid value F of the next auction item j; The calculation and bid value correction formula is: Where N is the total number of air defense weapon units, load i for No. WA i load, k is the coefficient for modifying the bid value F, F ij for WA i The revised bid value for auction item j; Step 3.11: After each auction contract, the WA with the largest task load i WA, which has the second largest workload for the same type of tasks j Share and communicate, if WA j Assign a task in its own task sequence to WA i If the exchange can increase the overall objective function value of both parties, the exchange contract mechanism will be triggered, and the exchange contract that increases the efficiency the most will be signed.

9. A task allocation system for edge air defense nodes based on genetic algorithm and contract network method, characterized in that: The system is used to implement the edge air defense node task allocation method based on genetic algorithm and contract network method as described in any one of claims 1 to 8. The system includes the first module to the third module, and the functions of each module are as follows: The first module is used to establish a target threat estimation and task allocation model; The second module uses an improved hybrid parthenogenetic algorithm to calculate the multi-platform collaborative task allocation problem; the improved hybrid parthenogenetic algorithm is designed through the genetic operator of integer coding + dynamic dual population, tournament selection + reverse crossover + adaptive mutation, combined with the elite pool update of the simulated annealing criterion and the mutation strategy triggered by concentration; The third module uses the improved contract network method to solve the task reallocation problem when the air defense platform encounters unexpected situations during mission execution; The improved contract network method realizes dynamic air defense task reallocation by extending the contract agreement process, introducing a multi-agent collaborative architecture, designing a consistent auction algorithm and dynamic bidding correction rules, and combining a load-sensitive contract exchange mechanism.

10. A mobile terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the edge air defense node task allocation method based on genetic algorithm and contract network method as described in any one of claims 1 to 8 is implemented.

Images