A multi-strategy improved distributed cross-domain collaborative task allocation optimization method

Abstract

The present invention provides a distributed cross-domain collaborative task allocation optimization method improved by multiple strategies, belonging to the field of planning and control technology. The method comprises the following steps: firstly, a target allocation optimization model is constructed based on three types of objective functions: comprehensive effectiveness, time cost and resource utilization; then, a control allocation optimization model is constructed based on three types of objective functions: sensor detection capability, transmission link connectivity and return link connectivity; then, a cross-domain collaborative task allocation optimization model is constructed by combining these two models and corresponding constraints. Finally, a distributed cross-domain collaborative task optimization allocation calculation is performed using a particle swarm algorithm improved by multiple strategies. The present invention adopts the above-mentioned distributed cross-domain collaborative task allocation optimization method improved by multiple strategies, which not only improves the overall combat effectiveness, but also enhances the concealment and coordination between combat nodes, provides strong technical support for cross-domain collaborative operations, and can effectively improve task execution efficiency and resource utilization efficiency.

Description

Technical Field

[0001] The present invention relates to the field of planning and control technology, and in particular to a distributed cross-domain collaborative task allocation optimization method with multi-strategy improvements. Background Art

[0002] Cross-domain collaborative task allocation involves the transfer of control rights and relay control of multi-domain nodes on land, sea and air. It optimizes the allocation of complex tasks by comprehensively utilizing the internal resources of the cluster on the basis of sharing situational awareness information.

[0003] At present, the task allocation optimization method mainly constructs the execution efficiency as the objective function, sets constraints and achieves the target optimization through the penalty function. In terms of optimization methods, it can be divided into traditional precise algorithms, heuristic algorithms and hybrid intelligent algorithms that combine multiple intelligent algorithms.

[0004] For example, in the 2023 journal article "Weapon Target Allocation Optimization Algorithm Based on Particle Swarm Genetic Taboo" published in the Journal of Zhengzhou University (Science Edition), the population diversity is regulated by introducing an adaptive selection ratio, and the crossover and mutation operators of the genetic algorithm and the neighborhood actions of the taboo search are improved to improve the convergence accuracy of the optimization algorithm. This method only addresses the problem of optimal allocation between ammunition and weapons, the task type is relatively simple, and does not involve cross-domain collaborative operations.

[0005] As in Example 2, in the 2022 Aeronaut journal paper "Multitarget allocation strategy based on adaptive SA-PSO algorithm", the advantages of simulated annealing algorithm and particle swarm algorithm were combined to propose a multi-objective optimization allocation strategy. This method is mainly aimed at the coordinated interception mission of multiple ground targets and does not involve cross-domain coordinated operations in the air and sea.

[0006] For example, in the 2024 Modern Manufacturing Engineering journal paper "Path Planning of Handling Robot Based on Improved Dung Beetle Optimization Algorithm", a heuristic swarm intelligence search algorithm was used to implement the robot path planning task. This method mainly optimizes and solves the path planning problem of small-scale unmanned swarms in a limited space, and does not involve complex tasks with multiple targets and types and cross-domain collaborative operations.

[0007] Since cross-domain collaborative task allocation has the characteristics of complex constraints, large scale and high timeliness, the above optimization algorithm has no better solution to the problem of unified allocation of multi-task objectives. Therefore, the present invention proposes a multi-strategy improved distributed cross-domain collaborative task optimization allocation method to solve the shortcomings of the above-mentioned prior art. Summary of the invention

[0008] The purpose of the present invention is to provide a multi-strategy improved distributed cross-domain collaborative task allocation optimization method, which not only improves the overall combat effectiveness, but also enhances the concealment and coordination between combat nodes, provides strong technical support for cross-domain collaborative combat, and can effectively improve task execution efficiency and resource utilization benefits.

[0009] To achieve the above object, the present invention provides a multi-strategy improved distributed cross-domain collaborative task allocation optimization method, comprising the following steps:

[0010] Step S1, constructing a target allocation optimization model based on three types of objective functions: comprehensive efficiency, time cost and resource utilization;

[0011] Step S2: construct a control allocation optimization model based on three objective functions: sensor detection capability, transmission link connectivity, and return link connectivity;

[0012] Step S3: Combining the target allocation optimization model and the control allocation optimization model, as well as corresponding constraints, to construct a cross-domain collaborative task allocation optimization model;

[0013] Step S4, using a multi-strategy improved particle swarm algorithm to perform distributed cross-domain collaborative task optimization allocation calculation;

[0014] Step S41, using the improved Fuch chaotic mapping algorithm to generate an initialization population to increase the balance of individual distribution;

[0015] Step S42, controlling whether to use the isolated point reverse evolution algorithm to update the population according to the decision number P, so as to expand the search space;

[0016] Step S43: using the optimal individual local perturbation evolutionary algorithm to update the population, so as to increase individual difference information and avoid falling into the local optimal solution;

[0017] Step S44: Update individual optimum and global optimum based on PSO algorithm, and update particle speed and position;

[0018] Step S45, iteratively execute steps S41 to S44 until the iteration termination criterion is met.

[0019] Preferably, in step S1, a target allocation optimization model is constructed based on three types of objective functions: comprehensive efficiency, time cost, and resource utilization, including:

[0020] (1);

[0021] (2);

[0022] in, represents the objective function of the target allocation optimization model; Indicates overall effectiveness, including target threat level , target value Probability of target being destroyed ; Indicates the time cost, including the target distance , Cluster movement speed ; Indicates resource utilization, including the actual number of resources allocated to the target , Target available allocation resource quantity ; , , All represent weighting coefficients;

[0023] Formula (2) is the combat domain node Assignment Target j The first row represents the combat domain nodes of all assigned targets. The total number does not exceed the total number of nodes in all combat domains d ; The second line represents each combat domain node Mapping associations to only one target j ; The third line represents each target j Maximum possible mappings m Nodes Quantity; the fourth line represents the combat domain node With the goal j The distance between Not exceeding the maximum theoretical range ; Represents the target-combat domain node mapping relationship function.

[0024] Preferably, in step S2, a control allocation optimization model is constructed based on three objective functions: sensor detection capability, transmission link connectivity, and return link connectivity, including:

[0025] (3);

[0026] (4);

[0027] in, represents the objective function of the control allocation optimization model; Indicates the sensor detection capability, including the total number of target identifications q , Total number of nodes in all combat domains d ; Indicates the connectivity of the sending link, including the combat domain node Total received power ; Indicates the connectivity of the backhaul link, including the aerial relay nodes Total received power ; Represents the connectivity function of the combat domain node-air relay node link;

[0028] Formula (4) is the aerial relay node Allocate control to combat domain nodes The first row represents all the combat domain nodes that are assigned control rights. The total number does not exceed the total number of nodes in all combat domains d ; The second line represents each combat domain node Only one aerial relay node is associated with the mapping ; The third row represents each aerial relay node k Maximum possible mappings Operational Domain Node quantity.

[0029] Preferably, in step S3, a cross-domain collaborative task allocation optimization model is constructed by combining the target allocation optimization model and the control allocation optimization model, as well as corresponding constraints, including:

[0030] (1);

[0031] (3);

[0032] (5).

[0033] Preferably, in step S41, an improved Fuch chaotic mapping algorithm is used to generate an initialization population to increase the balance of individual distribution, including:

[0034] (6);

[0035] in, represents the initial position of the population; Represents the population position after the improved Fuch chaotic map.

[0036] Preferably, in step S42, whether to use the isolated point reverse evolution algorithm to update the population is controlled according to the decision number P to expand the search space, including:

[0037] (7);

[0038] in, Indicates t +1 iteration of isolated points; Indicates t The isolated point of the iteration; represents the individual maximum extreme value; represents the individual minimum extreme value; represents the disturbance factor.

[0039] Preferably, in step S43, the optimal individual local perturbation evolutionary algorithm is used to update the population to increase individual difference information and avoid falling into a local optimal solution, including:

[0040] (8);

[0041] (9);

[0042] in, Indicates t +1 global optimal solution; Indicates t Sub-global optimal solution; represents the local worst solution; represents the global worst solution; represents the control coefficient; , Respectively represent the maximum and minimum values ​​of the control coefficient; Indicates no more than 10 -4 The constant of Represents the individual fitness function of particles.

[0043] Therefore, the present invention adopts the above-mentioned multi-strategy improved distributed cross-domain collaborative task allocation optimization method, and the beneficial technical effects are as follows:

[0044] (1) Construction of a large-scale complex task model: This model conducts correlation analysis and organic combination of multi-task target allocation and coordinated control between various combat domain nodes, accurately describing the multi-task optimization allocation problem in complex task scenarios, maximizing the concealment and coordination between cross-domain combat nodes, and improving overall combat effectiveness.

[0045] (2) A multi-strategy improved algorithm is proposed: It combines multiple optimization algorithm improvement strategies and comprehensively considers multiple constraint factors, accelerating the convergence of the algorithm while avoiding falling into the local optimal problem. It can be effectively applied to various cross-domain collaborative task allocation scenarios.

[0046] (3) Distributed planning structure design: Distributed task planning is used to improve the autonomous decision-making capabilities of each combat node, overcome the limitations of centralized decision-making, such as high communication requirements, high computational complexity, low fault tolerance, and poor system scalability, and fully mobilize all available resources for efficient task allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1The present invention provides a flowchart of a multi-strategy improved distributed cross-domain collaborative task allocation optimization method. DETAILED DESCRIPTION

[0048] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0049] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0050] Embodiment 1

[0051] like Figure 1 As shown, it is a flowchart of a multi-strategy improved distributed cross-domain collaborative task allocation optimization method of the present invention, which includes the following steps:

[0052] 1. Target allocation optimization model.

[0053] In order to optimize the allocation of multi-task objectives, three types of indicators, namely comprehensive effectiveness, time cost and resource utilization, can be used to evaluate and measure them. For the combined optimization problem of the above three types of indicators, a target allocation optimization model is constructed. First, the three types of objective functions are processed by the normalization method to unify the data scale and improve the convergence speed. Then, the values ​​of the three types of objective functions, namely comprehensive effectiveness, time cost and resource utilization, are weighted by the analytic hierarchy process (AHP). The specific optimization model and its constraints are shown in the following formula:

[0054] (1);

[0055] (2);

[0056] in, represents the objective function of the target allocation optimization model; Indicates overall effectiveness, including target threat level , target value Probability of target being destroyed ; Indicates the time cost, including the target distance , Cluster movement speed ; Indicates resource utilization, including the actual number of resources allocated to the target , Target available allocation resource quantity ; , , All represent weighting coefficients;

[0057] Formula (2) is the combat domain node Assignment Target jThe first row represents the combat domain nodes of all assigned targets. The total number does not exceed the total number of nodes in all combat domains d ; The second line represents each combat domain node Mapping associations to only one target j ; The third line represents each target j Maximum possible mappings m Nodes Quantity; the fourth line represents the combat domain node With the goal j The distance between Not exceeding the maximum theoretical range ; Represents the target-combat domain node mapping relationship function.

[0058] 2. Control allocation optimization model.

[0059] During cross-domain collaborative operations, command and control may be transferred or switched according to the surrounding situation or mission requirements. The specific switching timing is as follows: 1) When the land / sea combat node reaches the control switching distance, the air relay node can transfer control to the above node; 2) When the land / sea combat node receives serious threats or signal interference, control is transferred to the air relay node; 3) When a new combat target appears or a sudden situation occurs, control is transferred to the air relay node.

[0060] For the optimization problem of command and control rights allocated to combat domain nodes, three types of indicators can be used for evaluation and measurement: sensor detection capability, transmission link connectivity, and return link connectivity. For the optimization problem of the combination of the above three types of indicators, a control allocation optimization model is constructed. The specific optimization model and its constraints are shown in the following formula:

[0061] (3);

[0062] (4);

[0063] in, represents the objective function of the control allocation optimization model; Indicates the sensor detection capability, including the total number of target identifications q , Total number of nodes in all combat domains d ; Indicates the connectivity of the sending link, including the combat domain node Total received power ; Indicates the connectivity of the backhaul link, including the aerial relay nodes Total received power ; Represents the connectivity function of the combat domain node-air relay node link;

[0064] Formula (4) is the aerial relay node Allocate control to combat domain nodes The first row represents all the combat domain nodes that are assigned control rights. The total number does not exceed the total number of nodes in all combat domains d ; The second line represents each combat domain node Only one aerial relay node is associated with the mapping ; The third row represents each aerial relay node Maximum possible mappings Operational Domain Node quantity.

[0065] 3. Cross-domain collaborative task allocation optimization model.

[0066] For the problem of cross-domain collaborative task allocation, the control allocation optimization model is solved and calculated based on the target allocation optimization result, and the cross-domain collaborative task allocation optimization scheme is obtained. The corresponding optimization model and its constraints are a combination of the above two algorithms, as shown in the following formula:

[0067] (1);

[0068] (3);

[0069] (5).

[0070] 4. Multi-strategy improved distributed cross-domain collaborative task allocation optimization algorithm.

[0071] 1) PSO algorithm.

[0072] The particle swarm algorithm (PSO) is a random search algorithm that imitates the swarm intelligence of birds. The algorithm searches the space to record the global optimal position and speed found in each iteration, and determines whether to update the current optimal position and speed by setting the fitness of the objective function. Considered as particles in the particle swarm algorithm , the algorithm update formula is as follows:

[0073] (6);

[0074] (7);

[0075] in, Indicates t +1 iteration particle speed; Indicates t Iteration Particle speed; Indicates t +1 iteration particle location; Indicates t Iteration Particle location; Indicates passing t +1 search iteration particle The best historical position; Indicates passing t +1 search iteration to find the global optimal position of the entire population; Indicates the inertia weight, which is the effect of the previous moment's speed on the next movement; , represents the learning factor; represents the disturbance factor.

[0076] 2) Improve the Fuch chaotic mapping algorithm.

[0077] Since the initial population has a great influence on the optimization results of the entire algorithm, an improved Fuch chaotic mapping algorithm is used to increase the balance of individual distribution and avoid premature convergence of the algorithm. The algorithm formula is as follows:

[0078] (8);

[0079] in, represents the initial position of the population; Represents the population position after the improved Fuch chaotic map.

[0080] 3) Isolated point reverse evolution algorithm.

[0081] In order to further expand the algorithm's search space, an isolated point reverse evolution algorithm is proposed, which mainly performs reverse iterative evolution on isolated point individuals outside the population. The algorithm update formula is as follows:

[0082] (9);

[0083] in, Indicates t +1 iteration of isolated points; Indicates t The isolated point of the iteration; represents the individual maximum extreme value; Represents the individual minimum extreme value.

[0084] 4) Optimal individual local perturbation evolutionary algorithm.

[0085] In the process of optimization calculation, in order to increase individual difference information and avoid falling into the local optimal solution, an optimal individual local perturbation evolutionary algorithm is proposed. It mainly introduces the control coefficient factor to adaptively add local perturbations to the optimal individual, so as to obtain the updated global optimal solution.

[0086] (10);

[0087] (11);

[0088] in, Indicates t +1 global optimal solution; Indicates t Sub-global optimal solution; represents the local worst solution; represents the global worst solution; represents the control coefficient; , Respectively represent the maximum and minimum values ​​of the control coefficient; Indicates no more than 10 -4 The constant of Represents the individual fitness function of particles.

[0089] 5) Multi-strategy improvement of particle swarm optimization (PSO) algorithm process.

[0090] In order to achieve the optimal allocation of distributed cross-domain collaborative tasks, the particle swarm optimization algorithm is improved with multiple strategies to improve the speed and accuracy of solving complex problems. First, the improved Fuch chaotic mapping algorithm is introduced to generate the initial population to reduce the phenomenon of uneven population distribution or excessive individual differences; then the strategy of controlling probability is adopted, and the reverse evolution algorithm entering the extreme point is controlled by introducing the decision number P to expand the population search space; then the optimal individual local perturbation evolution algorithm is used to increase the individual difference information and avoid falling into the local optimal solution.

[0091] The specific steps of the algorithm are as follows:

[0092] Step 1: Set algorithm parameters for comprehensive performance , time cost , Resource Utilization , sensor detection capability , Send link connectivity , Backhaul Link Connectivity , and set the upper and lower bounds of the parameters according to the constraints of formula (5);

[0093] Step 2: According to formula (8), the initial population is Perform equalization processing;

[0094] Step 3: Construct two types of objective optimization functions according to formula (1) and formula (3) to calculate the fitness value of each particle ;

[0095] Step 4: If rand (0, 1) ≥ random decision number P (P is the number of iterations) t If the variable function increases and decreases), the algorithm falls into the local optimum and goes to step 5, otherwise jump to step 7;

[0096] Step 5: According to formula (9), use the isolated point reverse evolution algorithm to find the isolated point and update the population;

[0097] Step 6: According to formula (10) and formula (11), the optimal individual local perturbation evolutionary algorithm is used to obtain the current global optimal solution , to update the population;

[0098] Step 7: Update individual optimality based on PSO algorithm , global optimal ;

[0099] Step 8: Update particle velocity according to formula (6) and formula (7) ,Location ;

[0100] Step 9: Determine whether the iteration termination criteria are met. If the number of iterations reaches the maximum number of iterations, the algorithm stops; otherwise, return to step 3.

[0101] It is worth noting that the contents not elaborated in detail in the present invention are all prior art and are well known to those skilled in the art.

[0102] Therefore, the present invention adopts the above-mentioned multi-strategy improved distributed cross-domain collaborative task allocation optimization method, which not only improves the overall combat effectiveness, but also enhances the concealment and coordination between combat nodes, provides strong technical support for cross-domain collaborative combat, and can effectively improve task execution efficiency and resource utilization benefits.

[0103] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A multi-strategy improved distributed cross-domain collaborative task allocation optimization method, characterized in that: The following steps are involved: Step S1, constructing a target allocation optimization model based on three types of objective functions: comprehensive efficiency, time cost and resource utilization; Step S2: construct a control allocation optimization model based on three objective functions: sensor detection capability, transmission link connectivity, and return link connectivity; Step S3: Combining the target allocation optimization model and the control allocation optimization model, as well as corresponding constraints, to construct a cross-domain collaborative task allocation optimization model; Step S4, using a multi-strategy improved particle swarm algorithm to perform distributed cross-domain collaborative task optimization allocation calculation; Step S41, using the improved Fuch chaotic mapping algorithm to generate an initialization population; Step S42, controlling whether to adopt the isolated point reverse evolution algorithm to update the population according to the decision number P; Step S43, using the optimal individual local perturbation evolutionary algorithm to update the population; Step S44: Update individual optimum and global optimum based on PSO algorithm, and update particle speed and position; Step S45, iteratively executing steps S41 to S44 until the iteration termination criterion is met; In step S1, a target allocation optimization model is constructed based on three objective functions: comprehensive efficiency, time cost, and resource utilization, including: （1）； st （2）； in, represents the objective function of the target allocation optimization model; Indicates overall effectiveness, including target threat level , target value Probability of target being destroyed ; Indicates the time cost, including the target distance , Cluster movement speed ; Indicates resource utilization, including the actual number of resources allocated to the target , Target available allocation resource quantity ; , , All represent weighting coefficients; Formula (2) is the combat domain node Assignment Target The first row represents the combat domain nodes of all assigned targets. The total number does not exceed the total number of nodes in all combat domains ; The second line represents each combat domain node Mapping associations to only one target ; The third line represents each target Maximum possible mappings Nodes Quantity; the fourth line represents the combat domain node With the goal The distance between Not exceeding the maximum theoretical range ; Represents the target-combat domain node mapping relationship function; In step S2, based on three objective functions, namely, sensor detection capability, transmission link connectivity, and return link connectivity, a control allocation optimization model is constructed, including: （3）； （4）； in, represents the objective function of the control allocation optimization model; Indicates the sensor detection capability, including the total number of target identifications , Total number of nodes in all combat domains ; Indicates the connectivity of the sending link, including the combat domain node Total received power ; Indicates the connectivity of the backhaul link, including the aerial relay nodes Total received power ; Represents the connectivity function of the combat domain node-air relay node link; Formula (4) is the aerial relay node Allocate control to combat domain nodes The first row represents all the combat domain nodes that are assigned control rights. The total number does not exceed the total number of nodes in all combat domains ; The second line represents each combat domain node Only one aerial relay node is associated with the mapping ; The third row represents each aerial relay node Maximum possible mappings Operational Domain Node quantity; In step S3, a cross-domain collaborative task allocation optimization model is constructed by combining the target allocation optimization model and the control allocation optimization model, as well as corresponding constraints, including: （1）； （3）； st （5）。 2. According to the multi-strategy improved distributed cross-domain collaborative task allocation optimization method of claim 1, it is characterized in that: In step S41, an improved Fuch chaotic mapping algorithm is used to generate an initialization population, including: （6）； in, represents the initial position of the population; Represents the population position after the improved Fuch chaotic map.

3. According to the multi-strategy improved distributed cross-domain collaborative task allocation optimization method of claim 1, it is characterized in that: In step S42, whether to use the isolated point reverse evolution algorithm to update the population is controlled according to the decision number P, including: （7）； in, Indicates The isolated point of the iteration; Indicates t The isolated point of the iteration; represents the individual maximum extreme value; represents the individual minimum extreme value; represents the disturbance factor.

4. According to the multi-strategy improved distributed cross-domain collaborative task allocation optimization method of claim 1, it is characterized in that: In step S43, the population is updated using the optimal individual local perturbation evolutionary algorithm, including: （8）； （9）； in, Indicates Sub-global optimal solution; Indicates t Sub-global optimal solution; represents the local worst solution; represents the global worst solution; represents the control coefficient; , Respectively represent the maximum and minimum values ​​of the control coefficient; Indicates no more than The constant of Represents the individual fitness function of particles.

Images