Cluster clustering distribution optimization method and system based on coalition game

Abstract

The invention discloses a cluster clustering allocation optimization method and system based on a coalition game, and the method mainly comprises the steps: analyzing the task demands and resource constraints of a cluster clustering allocation scene, building a cluster clustering allocation mathematical model, and representing the coupling relation between the cluster clustering and allocation processes; carrying out cluster division distribution initialization; converting the cluster clustering distribution mathematical model into a clustering distribution alliance game solving model; and designing an alliance iterative optimization strategy based on the clustering allocation alliance game solving model, and generating a cluster clustering allocation optimization result through the alliance iterative optimization strategy. According to the method, the optimization problem of cluster clustering and distribution high coupling can be quickly solved, and the method is suitable for scenes such as communication resource distribution, target distribution and firepower distribution of large-scale clusters.

Description

Technical Field

[0001] The present invention belongs to the technical field of cluster command and control, and specifically relates to a cluster sub-clustering allocation optimization method and system based on alliance game. Background Art

[0002] Cluster task allocation technology has become a core support for fields such as industrial manufacturing and smart cities. For example, in the civilian sector, intelligent logistics robot clusters need to dynamically divide transport subgroups based on different order volumes to optimize delivery efficiency.

[0003] In order to achieve reasonable resource allocation and collaborative task execution, the cluster can assign different tasks to the agents in the cluster through clustering. For example, in a heterogeneous cluster, different amounts of resources are allocated according to the complexity of the task to improve the overall task execution efficiency of the cluster. In the face of sudden tasks or node failures, dynamic clustering can quickly reconstruct the cluster architecture to ensure the task completion rate. Although the cluster allocation optimization idea is very intuitive, due to the high coupling between the clustering process and the allocation process, it involves a series of problems such as the analysis of the coupling mechanism of clustering and allocation, and the coordinated optimization control of clustering and allocation. The current limitations of domestic research on cluster allocation optimization are mainly reflected in the following two aspects:

[0004] First, it's difficult to simultaneously achieve optimal dynamic clustering and task allocation. Traditional methods often employ static clustering or two-stage optimization (clustering first, then allocation), which decouples clustering and allocation. For example, in drone swarms, dynamic target position changes can cause clustering structures to fail, leading to lagging task allocation strategies and reduced overall resource utilization.

[0005] Second, it's difficult to ensure real-time clustering and allocation. Existing algorithms rely on offline computation or fixed-period updates, making them incapable of handling highly dynamic environments (such as surges in logistics demand). The high computational complexity of complex optimization models leads to delayed clustering and allocation responses, impacting task execution efficiency. Summary of the Invention

[0006] The purpose of the present invention is to provide a cluster allocation optimization method and system based on coalition game to solve the problems of unbalanced cluster allocation and task allocation, low optimization quality and computational efficiency in the current cluster allocation process.

[0007] The technical solutions for achieving the purpose of the present invention are:

[0008] A cluster allocation optimization method based on coalition game, comprising the following steps:

[0009] Analyze the task requirements and resource constraints of cluster allocation scenarios, establish a mathematical model for cluster allocation, and characterize the coupling relationship between cluster allocation and allocation processes;

[0010] Initialize cluster allocation;

[0011] Transform the mathematical model of cluster allocation into a cluster allocation alliance game solution model;

[0012] Based on the cluster allocation alliance game solution model, an alliance iterative optimization strategy is designed, and the cluster allocation optimization results are generated through the alliance iterative optimization strategy.

[0013] Furthermore, we analyze the task requirements and resource constraints of cluster allocation scenarios and establish a mathematical model for cluster allocation, which specifically includes:

[0014] Assume existence Tasks, use Cluster allocation is performed on the clusters composed of agents;

[0015] The utility function of each agent is established based on the weighted combination of the agent's energy consumption and task completion rate;

[0016] Based on the utility function, the benefit function of different agents collaborating to complete tasks after clustering is designed according to the collaboration achievement rate of each agent.

[0017] Based on the profit function, a mathematical model of cluster allocation is constructed.

[0018] Furthermore, the utility function of each agent is:

[0019]

[0020] in, For intelligent agents Task income; For intelligent agents Execute the task Energy consumption; For intelligent agents Execute the task completion rate; is the weight coefficient.

[0021] Furthermore, the profit function of different agents collaborating to complete tasks after clustering is:

[0022]

[0023] in, Representative cluster Task income; Cluster Any agent in Task income; Represents a cluster All agents in the Solution .

[0024] Furthermore, the nonlinear function for:

[0025]

[0026] in, The collaboration achievement rate.

[0027] Furthermore, the cluster allocation mathematical model is:

[0028]

[0029] Where, represents the function used to solve the total benefits of all clusters, is the number of clusters; is the decision variable related to cluster assignment, indicating the cluster Whether to execute the task ,satisfy

[0030] ;

[0031] Indicates that an agent can only join one cluster, , both are positive integers, representing the identity of the cluster, and Respectively represent Clusters and clusters; Indicates that all agents must participate in the clustering process and join a cluster; Indicates that each task can only be executed by one cluster; Indicates that each cluster executes at most one task; Indicates that all tasks must be executed.

[0032] Furthermore, the cluster allocation initialization specifically includes: each agent in the initial cluster will be assigned a task based on its maximum benefit To determine the task to be performed, if several agents choose the same task, they will form a cluster together. , and calculate the task assigned to the cluster Profit .

[0033] Furthermore, the cluster allocation alliance game solution model is:

[0034]

[0035] Where, , , indicating that the same agent cannot join different clusters. In the cluster allocation coalition game solution model, clusters are alliance partitions; Indicates that any agent must join the cluster; Indicates that the number of stable alliance partitions formed is not less than the number of tasks, , representing a cluster Assign tasks of income.

[0036] Furthermore, the alliance iterative optimization strategy is:

[0037] Each agent has its own task Maximum benefit joins the initial alliance division Afterwards, each agent By continuously performing the three switching operations of leaving the alliance, remaining unchanged, and switching alliances, the alliance structure is updated to form new alliance partitions, which gradually stabilizes the total alliance revenue in the direction of increasing, and finally obtains a stable alliance partition, that is, the total alliance revenue the largest league division; Indicates the clusters.

[0038] A cluster allocation optimization system based on coalition game, comprising:

[0039] Cluster allocation mathematical model construction unit, which analyzes the task requirements and resource constraints of cluster allocation scenarios, is used to establish a cluster allocation mathematical model, and characterizes the coupling relationship between cluster allocation and allocation processes;

[0040] Initialization unit, used for initialization of cluster allocation;

[0041] A cluster allocation mathematical model conversion unit, used to convert the cluster allocation mathematical model into a cluster allocation alliance game solution model;

[0042] The iterative solution unit is based on the cluster allocation alliance game solution model and is used to design the alliance iterative optimization strategy, and generate the cluster allocation optimization results through the alliance iterative optimization strategy.

[0043] Compared with the existing technology, the present invention has the following advantages: 1) It analytically characterizes the coupling relationship between task allocation and agent clustering process, solves the clustering and allocation problems at the same time, and improves computational efficiency; 2) The benefits corresponding to each cluster composed of different participants are calculated in advance, and the benefits can be updated in real time during the cluster allocation process without time-consuming calculations, thereby enhancing the real-time nature of the cluster allocation result generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the principle of the present invention.

[0045] Figure 2 It is a schematic diagram of an application scenario of the present invention. DETAILED DESCRIPTION

[0046] In order to better illustrate the purpose and function of the present invention, the present invention is further described in detail below with reference to specific examples.

[0047] Aiming at the problems of imbalanced clustering and task allocation, low optimization quality and computational efficiency in the current cluster allocation process, the present invention discloses a cluster allocation optimization method that supports dynamic task scenarios. The principle diagram is shown in the attached figure. Figure 1 This method controls the collaborative relationship between intelligent agents through coalition games and completes task allocation at the same time, achieving efficient resource utilization and balanced task allocation in dynamic scenarios. The specific steps are as follows:

[0048] Firstly, the task requirements and resource constraints of cluster allocation scenarios are analyzed, and a mathematical model of cluster allocation is established.

[0049] Assume existence Tasks, use The cluster composed of agents is divided into clusters and allocated. The specific scenario diagram is as follows Figure 2 As shown. The utility function of each agent is established based on the weighted combination of the agent's energy consumption and task completion rate:

[0050]

[0051] in, For intelligent agents Task income; For intelligent agents Execute the task Energy consumption; For intelligent agents Execute the task completion rate; is the weight coefficient. Then, according to the collaboration achievement rate of each agent, the profit function of different agents in clustering to complete the task collaboratively is designed:

[0052]

[0053] in, Representative cluster Task income; Cluster Any agent in Task income; Represents a cluster All agents in the Solution , the special case of which is , but not limited to this form, in this special case is the collaboration achievement rate, For the agents in the cluster to perform tasks The constructed cluster allocation mathematical model is as follows:

[0054]

[0055] Where, is the number of clusters; is the decision variable related to cluster assignment, indicating the cluster Whether to execute the task ,satisfy

[0056] ;

[0057] Indicates that an agent can only join one cluster; Indicates that all agents must participate in the clustering process and join a cluster; Indicates that each task can only be executed by one cluster; Indicates that each cluster executes at most one task; Indicates that all tasks must be executed.

[0058] This step constructs a mathematical model of cluster sub-clustering and allocation for scenarios such as large-scale cluster communication resource allocation, target allocation, and firepower allocation. It analytically characterizes the coupling relationship between cluster sub-clustering and allocation processes, and can solve clustering and allocation problems at the same time.

[0059] Secondly, the initial clustering allocation is performed on the agents in the cluster. The initial clustering allocation is an important step in the clustering allocation optimization method. Each agent in the initial cluster will be assigned according to its maximum benefit for a task. To determine the task to be performed, if several agents choose the same task, they will form a cluster together. , and calculate the task assigned to the cluster according to the profit function formula of the present invention Profit .

[0060] In this step, an initial cluster allocation strategy is designed to accelerate the solution of cluster allocation optimization problem, and the cluster allocation model is transformed into a coalition game solution model, providing a new idea for the reduction solution of the strong coupling problem of cluster allocation.

[0061] Next, the constructed cluster allocation mathematical model will be transformed into a coalition game model based on the principle of coalition game. Coalition game, a type of cooperative game, refers to the process in which game participants form a stable alliance with other participants in the form of alliance and cooperation. Game refers to a situation in which several rational participants produce good or bad results through complex interactions; alliance refers to the situation in which all rational participants have the intention to cooperate and hope to produce better results by forming a cooperative organization. Therefore, the main problem to be solved by coalition game is how to form an appropriate cooperative organization to achieve the desired results. Coalition game contains two elements. One is the set of participants who need to form an alliance. , forming an alliance according to the rules and constraints of the specific model, followed by alliance benefits, which can evaluate the size of the alliance benefits. Therefore, the mathematical model of the alliance game can be defined as .

[0062] From the above definition, we can see that the problem to be solved in the alliance game is the same as that in the cluster allocation problem. Individual agents are considered participants in a coalition game, and the clusters they form are considered coalitions of these participants. Solving the cluster allocation problem strategy can be transformed into determining the coalition structure. This allows the cluster allocation process to be equivalent to the splitting or disintegration of a coalition, ultimately forming multiple disjoint coalitions through repeated selection cycles. Therefore, coalition games are feasible for solving the cluster allocation problem in multi-agent systems.

[0063] The multiple disjoint coalitions formed by agents are called coalition partitions, or clusters. The set of coalition game participants composed of agents is , then the union partition is defined as the set ,in , the alliance in the set П satisfy and .

[0064] From the mathematical model of alliance game, we can know that alliance game includes two elements: the set of participants that form the alliance and alliance benefits to assess alliance value , the coalition game model is available Indicates the number of tasks. Determine the initial alliance task allocation , each league division Assigned to specific tasks , its benefits are The goal of the game is to form a stable alliance partition and make the total benefits of all alliances Maximization, the transformed alliance game model is as follows:

[0065]

[0066] Where, , , indicating that the same agent cannot join different alliance partitions; Indicates that any agent must join the alliance partition; Indicates that the number of stable alliance partitions formed is not less than the number of tasks.

[0067] Finally, the iterative optimization strategy design of alliance partition is carried out, and it is gradually iterated until a stable alliance partition is formed and the total benefits of all alliances are maximized. Maximization can quickly generate cluster allocation optimization results. Based on the corresponding alliance game mathematical model and the initial alliance partition task allocation, the initial alliance agent members are not necessarily the optimal composition, so it is necessary to continuously switch alliances through split strategies to obtain stable alliance partitions. The specific split strategy definition is given below. Each agent is based on its own task Maximum benefit joins the initial alliance division Afterwards, each agent The alliance structure will be updated by continuously leaving the alliance, remaining unchanged, and switching alliances to form new alliance partitions. The three behaviors of updating the alliance partitions are called switching operations, that is, the iterative strategy of the multi-agent cluster allocation problem. The goal is to gradually stabilize the total income of the alliance in the direction of increase by continuously performing switching operations, and finally obtain a stable alliance partition.

[0068] Each agent in this alliance game model There are three switching operations that can be defined to express ,in Indicates the alliance partition Agents within Leave the current alliance; Indicates the alliance partition Agents within Remain in the current alliance unchanged; Indicates the alliance partition Agents within Leave the current alliance and join another one .

[0069] From the above analysis, we can see that in order to achieve the split goal, we first establish the corresponding initial alliance partition according to the number of tasks. Each agent joins the corresponding initial alliance according to its maximum benefit for each task. After that, each agent will continue to switch, and the benefits of all alliance partitions will be calculated once each switching operation is completed. , if it is greater than the previous alliance total income, then update the alliance partition. The increasing direction is selected several times until When it no longer changes, stop the iteration. At the same time, in order to make the results converge quickly, Calculations are simplified to predefined table queries, which are significantly faster than traditional dynamic programming or heuristic algorithms, improving computing efficiency.

[0070] The algorithm flow of the cluster allocation optimization method is as follows: first, establish the initial alliance partition according to the number of tasks and assign tasks according to the maximum benefit of each agent. Agents with the same tasks join the corresponding alliance and calculate the total benefit of the alliance at this time. ; Then iterate from agent 1, and iterate from agent 1 to Perform three switching operations: Leaving the Alliance: ; Remain unchanged: ; Switch alliance: ; The total alliance revenue updated by the three operations is recorded as , , , take the maximum value of the three Total alliance revenue compared to the previous one Compare, if it increases, update the alliance partition and the total alliance income according to the corresponding operation, otherwise keep the original alliance partition unchanged; after the update is completed, judge the next agent until the completion of the first The judgment of each agent is regarded as the completion of a round of alliance game. If multiple rounds of alliance game are carried out and the total profit does not change, it means that the alliance partition composition has reached a relatively stable state, and the algorithm is completed at this time.

[0071] A stable alliance partition refers to a situation in which each participant in a coalition game joins the optimal alliance that can achieve the goal of the game. If any participant in the formed alliance partition switches its alliance alone, its profit will not increase. The alliance partition at this time is called a stable alliance partition. According to analysis, after continuous iteration, each agent as a coalition participant joins the optimal alliance that can make the total profit of the alliance The largest alliance , the membership of each alliance gradually stabilizes, and the alliance partition formed at this time is stable. Because if the alliance partition is not stable at this time, it means that there is a switching operation that causes the members in the alliance partition to continue to trigger switching operations. These two operations are contradictory. Therefore, the alliance partition formed in the end must be a stable alliance partition.

[0072] Based on the method, the present invention also provides a cluster allocation optimization system based on alliance game, including:

[0073] Cluster allocation mathematical model construction unit, which analyzes the task requirements and resource constraints of cluster allocation scenarios, is used to establish a cluster allocation mathematical model, and characterizes the coupling relationship between cluster allocation and allocation processes;

[0074] Initialization unit, used for initialization of cluster allocation;

[0075] A cluster allocation mathematical model conversion unit, used to convert the cluster allocation mathematical model into a cluster allocation alliance game solution model;

[0076] The iterative solution unit is based on the cluster allocation alliance game solution model and is used to design the alliance iterative optimization strategy, and generate the cluster allocation optimization results through the alliance iterative optimization strategy.

[0077] Example

[0078] Assume that in a certain scenario, eight drones need to perform four tasks. Based on the model description, clusters A, B, C, and D can be assigned to these four tasks. The multi-agent clustering strategy based on coalition game can be divided into two stages. In the first stage, the four tasks A, B, C, and D are initially assigned according to the maximum payoff of each drone. Drones assigned the same task are then assigned to the corresponding cluster, with each cluster corresponding to a coalition.

[0079] In order to achieve rapid cluster allocation, the corresponding benefits of the alliance that each drone may form are given according to the profit function formula, as shown in Table 1.

[0080] Table 1. The corresponding income table of the alliance with different members of the present invention

[0081]

[0082] In the process of switching alliance iterations, the process of repeatedly calculating alliance benefits is simplified to table lookup, without repeatedly calling the benefit function, which improves the efficiency of program execution. In order to quickly determine the alliance members in the program and obtain the corresponding alliance benefits, the number of rows in Table 1 corresponding to the benefits of different drones is calculated according to the formula Arrangement, where n is the drone number , and the columns are the different benefits for each task. For example, the benefits of each task formed by the alliance of agents 2, 3, and 5 are in the corresponding rows of Table 1. .

[0083] Step 1: Analyze the case study and establish a mathematical model for drone cluster allocation.

[0084]

[0085] Where: is the decision variable related to task allocation, representing the cluster Whether to execute the task , ; Indicates that a drone can only join one cluster; Indicates that all drones must participate in the clustering process to join the cluster; Indicates that each task can only be executed by one cluster; Indicates that each cluster executes at most one task; Indicates that all tasks must be executed.

[0086] Step 2: Perform initial cluster allocation. Initially allocate the drones according to their maximum benefits. The results are shown in Table 2. From Table 2, we can see that the alliance that performs task A is , the alliance that performs task B is , the alliance that performs task C is , the alliance that performs task D is , corresponding to the income They are 6, 8, 9, and 8 respectively. The total profit of the alliance is 31.

[0087] Table 2 Initial task allocation table of the present invention example

[0088]

[0089] Step 3: Convert the mathematical model of drone cluster allocation into a coalition game solution model.

[0090]

[0091] Where, , , indicating that the same drone cannot join different alliance partitions; Indicates that any drone must join the alliance division; Indicates that the number of stable alliance partitions formed is not less than the number of tasks.

[0092] Step 4: Run the alliance iterative optimization strategy to generate the cluster allocation optimization results. This means that switching operations are continuously performed on a drone-by-drone basis. Each drone must select the switching operation that maximizes the total benefit of all alliances and update the alliance partition accordingly.

[0093] (1) The switching operation of UAV 1 is shown in Table 3. , update each alliance partition and total revenue.

[0094] Table 3 Schematic diagram of switching operation of UAV 1

[0095]

[0096] (2) The switching operation for UAV 2 is shown in Table 4. , update each alliance partition and total revenue.

[0097] Table 4 Schematic diagram of switching operation of UAV 2

[0098]

[0099] (3) The switching operation of UAV 3 is shown in Table 5. , update each alliance partition and total revenue.

[0100] Table 5 Schematic diagram of switching operation of UAV 3

[0101]

[0102] (4) The switching operation of UAV 4 is shown in Table 6. , update each alliance partition and total revenue.

[0103] Table 6 Schematic diagram of switching operation of drone 4

[0104]

[0105] (5) The switching operation of UAV 5 is shown in Table 7. , update each alliance partition and total revenue.

[0106] Table 7 Schematic diagram of switching operation of drone 5

[0107]

[0108] (6) The switching operation of UAV 6 is shown in Table 8. , update each alliance partition and total revenue.

[0109] Table 8 Schematic diagram of switching operation of UAV 6

[0110]

[0111] (7) The switching operation of UAV 7 is shown in Table 9. , update each alliance partition and total revenue.

[0112] Table 9 Schematic diagram of switching operation of UAV 7

[0113]

[0114] (8) The switching operation of UAV 8 is shown in Table 10. , update each alliance partition and total revenue.

[0115] Table 10 Schematic diagram of UAV 8 switching operation

[0116]

[0117] The alliance partitions and corresponding benefits after the first iteration are shown in Table 11.

[0118] Table 11 Optimization results of the first iteration of the present invention

[0119]

[0120] To determine the total revenue of the alliance To determine whether the solution remains unchanged over multiple iterations, multiple rounds of selection are performed. The optimization results after the second iteration are shown in Table 12, and the optimization results after the third iteration are shown in Table 13. Analysis shows that after the first iteration, the alliance partitions and total alliance revenue remain optimal and unchanged in subsequent rounds of selection. This indicates that the alliance game has achieved a relatively optimal solution and formed a stable alliance partition.

[0121] Table 12 Optimization results of the second iteration of the present invention

[0122]

[0123] Table 13 Optimization results of the third iteration of the present invention

[0124]

[0125] Step 5: Obtain stable alliance partitions. The calculated allocations that maximize the total alliance benefit and avoid conflicts are shown in Table 14. This completes the optimization of the cluster allocation of the drone cluster in the example using alliance game theory.

[0126] Table 14 Final optimization results of the example of the present invention

[0127]

[0128] The optimization results show that the present invention can quickly solve the optimization problem of highly coupled clustering and allocation, and is suitable for scenarios such as communication resource allocation, target allocation, and firepower allocation of large-scale clusters.

[0129] Although the above describes the specific embodiments of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without any creative work are still within the scope of protection of the present invention.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A cluster allocation optimization method based on coalition game, characterized in that: Including steps: Analyze the task requirements and resource constraints of cluster allocation scenarios, establish a mathematical model for cluster allocation, and characterize the coupling relationship between cluster allocation and allocation processes; Initialize cluster allocation; Transform the mathematical model of cluster allocation into a cluster allocation alliance game solution model; Based on the cluster allocation alliance game solution model, an alliance iterative optimization strategy is designed, and the cluster allocation optimization results are generated through the alliance iterative optimization strategy.

2. The cluster allocation optimization method based on alliance game according to claim 1 is characterized in that: Analyze the task requirements and resource constraints of cluster allocation scenarios and establish a mathematical model for cluster allocation, including: Assume existence Tasks, use Cluster allocation is performed on the clusters composed of agents; The utility function of each agent is established based on the weighted combination of the agent's energy consumption and task completion rate; Based on the utility function, the benefit function of different agents collaborating to complete tasks after clustering is designed according to the collaboration achievement rate of each agent. Based on the profit function, a mathematical model of cluster allocation is constructed.

3. The cluster allocation optimization method based on alliance game according to claim 2 is characterized in that: The utility function of each agent is: ； in, For intelligent agents Task income; For intelligent agents Execute the task Energy consumption; For intelligent agents Execute the task completion rate; is the weight coefficient.

4. The cluster allocation optimization method based on alliance game according to claim 3 is characterized in that: The profit function of different agents collaborating to complete the task after clustering is: ； in, Representative cluster Task income; Cluster Any agent in Task income; Represents a cluster All agents in the Solution .

5. The cluster allocation optimization method based on alliance game according to claim 4 is characterized in that: The nonlinear function for: ； in, The collaboration achievement rate.

6. The cluster allocation optimization method based on alliance game according to claim 4 is characterized in that: The cluster allocation mathematical model is: ； Where, represents the function used to solve the total benefits of all clusters, is the number of clusters; is the decision variable related to cluster assignment, indicating the cluster Whether to execute the task ,satisfy: ； Indicates that an agent can only join one cluster, , both are positive integers, representing the identity of the cluster, and Respectively represent Clusters and clusters; Indicates that all agents must participate in the clustering process and join a cluster; Indicates that each task can only be executed by one cluster; Indicates that each cluster executes at most one task; Indicates that all tasks must be executed.

7. The cluster allocation optimization method based on alliance game according to claim 6, characterized in that: The cluster allocation initialization specifically includes: each agent in the initial cluster will be assigned according to its maximum benefit for a task To determine the task to be performed, if several agents choose the same task, they will form a cluster together. , and calculate the task assigned to the cluster Profit .

8. The cluster allocation optimization method based on alliance game according to claim 6 is characterized in that: The cluster allocation alliance game solution model is: ； Where, , , indicating that the same agent cannot join different clusters. In the cluster allocation coalition game solution model, clusters are alliance partitions; Indicates that any agent must join the cluster; Indicates that the number of stable alliance partitions formed is not less than the number of tasks, , representing a cluster Assign tasks of income.

9. The cluster allocation optimization method based on alliance game according to claim 2, characterized in that: The alliance iterative optimization strategy is: Each agent has its own task Maximum benefit joins the initial alliance division Afterwards, each agent By continuously performing the three switching operations of leaving the alliance, remaining unchanged, and switching alliances, the alliance structure is updated to form new alliance partitions, which gradually stabilizes the total alliance revenue in the direction of increasing, and finally obtains a stable alliance partition, that is, the total alliance revenue the largest league division; Indicates the clusters.

10. A cluster allocation optimization system based on alliance game for implementing any of the methods described in claims 1-9, characterized in that: include: Cluster allocation mathematical model construction unit, which analyzes the task requirements and resource constraints of cluster allocation scenarios, is used to establish a cluster allocation mathematical model, and characterizes the coupling relationship between cluster allocation and allocation processes; Initialization unit, used for initialization of cluster allocation; A cluster allocation mathematical model conversion unit, used to convert the cluster allocation mathematical model into a cluster allocation alliance game solution model; The iterative solution unit is based on the cluster allocation alliance game solution model and is used to design the alliance iterative optimization strategy, and generate the cluster allocation optimization results through the alliance iterative optimization strategy.

Images