A method and system for modeling and collaborative optimization of UAV swarm process behavior
By combining top-down decomposition with BPNN-NSGA-II, the problem of suboptimal top-level design in UAV swarm behavior optimization is solved, the global optimal solution is determined, and the collaborative optimization capability of UAV swarms is improved.
Patent Information
- Application Number
- CN202111654848.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-30
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2041-12-30
AI Technical Summary
During the self-organization process, drone swarms suffer from problems such as individual decision-making leading to loose swarm organization, unsatisfactory top-level design results, lack of quantitative characterization in existing methods, and difficulty in obtaining the global optimal solution.
A top-down, layer-by-layer decomposition strategy was adopted. Sample data was obtained using the NetLogo simulation platform. A function mapping relationship was established through a backpropagation neural network (BPNN). Multi-objective optimization was performed by combining the second-generation non-dominated sorting genetic algorithm (NSGA-Ⅱ) to determine the global optimal solution.
It enables effective prediction of drone swarm behavior, avoids decision-making getting stuck in local optima, explores the underlying optimal decision parameters that affect top-level design, and improves the accuracy and efficiency of swarm collaborative optimization.
Smart Images

Figure CN114492735B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of group intelligence behavior, and in particular to a method and system for modeling and collaborative optimization of unmanned aerial vehicle cluster process behavior. Background Art
[0002] A drone swarm is composed of a certain number of homogeneous and heterogeneous drones, which use sensors to exchange information and make real-time adjustments through feedback to achieve designated tasks. Compared with single drones, drone swarms involve larger scales, higher coordination, and stronger intelligence. Nowadays, drone swarm technology has been widely used in national defense construction and civilian fields. Its research focuses on how to model the process behavior of drone swarms and optimize the coordination efficiency, and have the ability to adapt to complex, changing, and uncertain environments.
[0003] Due to the emergence of self-organization of drone swarms, individual decisions can easily lead to loose and unplanned cluster organization, resulting in unsatisfactory top-level design results. Therefore, it is necessary to optimize the process behavior of drone swarms. However, the primary problem in achieving collaborative optimization of drone swarm behavior is how to adjust the underlying parameters to achieve the optimal design of the top level. In addition to considering the individual properties of a single machine, the optimization of drone swarm process behavior also needs to consider the cluster scale, spatial structure, and coordination mode. It can essentially be regarded as a multi-objective optimization design problem.
[0004] As the number of drones increases and the action states overlap, there are often problems such as decision coupling, complex parameters, and unclear constraints, which ultimately make it difficult to obtain the global optimal solution for the optimization of drone cluster process behavior. Therefore, in order to explore the optimal design of drone clusters, scholars at home and abroad have conducted relevant research, effectively achieved local optimization, and proposed corresponding optimization strategies. However, the existing methods have not conducted in-depth research from the perspective of "individual-group-top" in the emergence process, and the optimal design of drone clusters also lacks quantitative representation. Summary of the invention
[0005] The purpose of the present invention is to provide a method and system for modeling and collaborative optimization of the process behavior of a drone swarm, which can effectively realize the behavior prediction of a drone swarm, avoid the decision-making falling into local optimality, and help explore the underlying optimal decision parameters that affect the top-level design.
[0006] To achieve the above object, the present invention provides the following solutions:
[0007] A UAV swarm process behavior modeling and collaborative optimization method, comprising:
[0008] The drone swarm system is decomposed layer by layer from top to bottom to determine the individual decision-making layer, group behavior layer and top-level design layer; the parameters of the individual decision-making layer include: attack capability, defense capability, mobility, communication capability and obstacle avoidance capability; the parameters of the group behavior layer include: group size, group density, clustering coefficient and average distance; the parameters of the top-level design layer include: task completion, design cost and design efficiency;
[0009] The NetLogo simulation platform is used to obtain sample data of the drone cluster system; the sample data includes: parameters of the individual decision-making layer, parameters of the group behavior layer, and parameters of the top-level design layer;
[0010] The back propagation neural network (BPNN) is used to establish the functional mapping relationship between individual decision-making and group behavior, and the functional mapping relationship between group behavior and top-level design.
[0011] Taking the task completion, design cost and design efficiency in the parameters of the top-level design layer as the optimization goals, and the parameter range of the individual decision layer as the constraint condition, the objective function of multi-objective optimization is constructed based on the functional mapping relationship between individual decision-making and group behavior and the functional mapping relationship between group behavior and top-level design;
[0012] Based on the second generation non-dominated sorting genetic algorithm (NSGA-Ⅱ), a global search is performed on the objective function of multi-objective optimization to determine the global optimal solution.
[0013] Optionally, the method of obtaining sample data of the drone cluster system by using the NetLogo simulation platform further includes:
[0014] The sample data is normalized.
[0015] Optionally, the use of a back propagation neural network (BPNN) to respectively establish a functional mapping relationship between individual decision-making and group behavior and a functional mapping relationship between group behavior and top-level design specifically includes:
[0016] The parameters of the individual decision-making layer are used as the input of BPNN, and the parameters of the group behavior layer are used as the output of BPNN to establish the functional mapping relationship between individual decision-making and group behavior.
[0017] The parameters of the group behavior layer are used as the input of BPNN, and the parameters of the top-level design layer are used as the output of BPNN to establish a functional mapping relationship between individual decision-making and group behavior.
[0018] Optionally, the use of a back propagation neural network (BPNN) to respectively establish a functional mapping relationship between individual decision-making and group behavior and a functional mapping relationship between group behavior and top-level design specifically includes:
[0019] Using the formula Determine the input-output relationship of BPNN;
[0020] Among them, x i Indicates the input value; y k Indicates expected value; Represents the weight between input layer i and hidden layer j; represents the weight between hidden layer j and output layer k; a j represents the threshold of hidden layer j; b k represents the threshold of output layer k, n represents the number of hidden layer nodes, m represents the number of input layer nodes, and k represents the number of output layer nodes.
[0021] Optionally, the task completion, design cost and design efficiency in the parameters of the top-level design layer are taken as optimization targets, the parameter range of the individual decision layer is taken as a constraint condition, and the objective function of multi-objective optimization is constructed based on the functional mapping relationship between individual decision-making and group behavior and the functional mapping relationship between group behavior and top-level design, specifically including:
[0022] Using the formula Construct objective functions for multi-objective optimization;
[0023] Among them, Z1 represents the task completion, Z2 represents the design cost, Z3 represents the design efficiency, and p and q are constants.
[0024] A UAV swarm process behavior modeling and collaborative optimization system, comprising:
[0025] The hierarchical decomposition module is used to hierarchically decompose the drone swarm system using a top-down layer-by-layer decomposition strategy to determine the individual decision-making layer, group behavior layer and top-level design layer; the parameters of the individual decision-making layer include: attack capability, defense capability, mobility capability, communication capability and obstacle avoidance capability; the parameters of the group behavior layer include: group size, group density, clustering coefficient and average distance; the parameters of the top-level design layer include: task completion, design cost and design efficiency;
[0026] The sample data acquisition module is used to obtain sample data of the drone cluster system using the NetLogo simulation platform; the sample data includes: parameters of the individual decision-making layer, parameters of the group behavior layer, and parameters of the top-level design layer;
[0027] A function mapping relationship establishment module is used to establish the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design by using a back propagation neural network (BPNN);
[0028] The objective function construction module of multi-objective optimization is used to construct the objective function of multi-objective optimization based on the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design, taking the task completion, design cost and design efficiency in the parameters of the top-level design layer as the optimization objectives and the parameter range of the individual decision-making layer as the constraint condition;
[0029] The global optimal solution determination module is used to perform a global search for the objective function of multi-objective optimization based on the second-generation non-dominated sorting genetic algorithm NSGA-Ⅱ to determine the global optimal solution.
[0030] Optionally, it also includes:
[0031] The data normalization module is used to perform normalization processing on the sample data.
[0032] Optionally, the function mapping relationship establishing module specifically includes:
[0033] The first function mapping relationship establishment unit is used to establish the function mapping relationship between individual decision-making and group behavior by taking the parameters of the individual decision-making layer as the input of BPNN and taking the parameters of the group behavior layer as the output of BPNN;
[0034] The second function mapping relationship establishing unit is used to establish the function mapping relationship between individual decision-making and group behavior by taking the parameters of the group behavior layer as the input of BPNN and taking the parameters of the top design layer as the output of BPNN.
[0035] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:
[0036] The method and system for modeling and collaborative optimization of the process behavior of drone clusters provided by the present invention obtains cluster data from the NetLogo platform and decomposes it layer by layer from top to bottom, and constructs the function mapping relationship between adjacent layers after decomposition based on BPNN; on this basis, the NSGA-Ⅱ algorithm is used to realize layer-by-layer global search. Compared with other single intention recognition methods currently available, the present invention can effectively realize the behavior prediction of drone clusters, avoid the decision-making falling into the local optimum, and help explore the bottom-level optimal decision parameters that affect the top-level design. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0038] Figure 1 A schematic diagram of a process behavior modeling and collaborative optimization method for a drone swarm provided by the present invention;
[0039] Figure 2 It is a schematic diagram of the overall research route of UAV swarm process behavior modeling and collaborative optimization method based on BPNN and NSGA-Ⅱ;
[0040] Figure 3 Construct a model framework for strategies based on top-down hierarchical decomposition;
[0041] Figure 4 The BPNN structure diagram used is shown in Figure 2.
[0042] Figure 5 This is the NetLogo simulation platform diagram used to obtain data;
[0043] Figure 6 This is the fitting effect diagram of the real value and predicted value of "Group Behavior-Top Design";
[0044] Figure 7 This is the fitting effect diagram of the real value and predicted value of "Group Behavior-Top Design";
[0045] Figure 8 Visualization of the Pareto solution set optimized for a swarm of drones.
[0046] Fig. 9 This is a schematic diagram of the structure of a UAV cluster process behavior modeling and collaborative optimization system provided by the present invention. DETAILED DESCRIPTION
[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0048] The purpose of the present invention is to provide a method and system for modeling and collaborative optimization of the process behavior of a drone swarm, which can effectively realize the behavior prediction of a drone swarm, avoid the decision-making falling into local optimality, and help explore the underlying optimal decision parameters that affect the top-level design.
[0049] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0050] Figure 1 This is a flow chart of a method for modeling and collaborative optimization of UAV cluster process behavior provided by the present invention. Figure 2 The overall research route diagram of UAV swarm process behavior modeling and collaborative optimization method based on BPNN and NSGA-Ⅱ is shown in the figure. Figure 1 and Figure 2 As shown, the present invention provides a UAV cluster process behavior modeling and collaborative optimization method, including:
[0051] S101, the drone cluster system is decomposed layer by layer using a top-down decomposition strategy to determine the individual decision-making layer, group behavior layer and top-level design layer, and Figure 3 As shown; the parameters of the individual decision-making layer include: attack capability, defense capability, mobility, communication capability and obstacle avoidance capability; the parameters of the group behavior layer include: group size, group density, clustering coefficient and average distance; the parameters of the top-level design layer include: task completion, design cost and design efficiency; that is, the top level is equivalent to the command and control center, which records the overall status of the unmanned cluster.
[0052] S102, using the NetLogo simulation platform to obtain sample data of the drone cluster system; the sample data includes: the parameters of the individual decision-making layer X = [x1; x2; ...; x n ], the parameters of the group behavior layer Y = [y1; y2; ...; y n ] and the top-level design layer parameters Z = [z1; z2; ...; z n ];
[0053] In order to increase data diversity, improve the accuracy of model training and the final prediction effect, S102 and later also include:
[0054] The sample data is normalized.
[0055] The data X and Y are normalized between [0,1], and Z is normalized between [-1,1].
[0056] S103, using the nonlinear regression capability of the back propagation neural network BPNN (such as Figure 4 As shown in the figure, the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design are established respectively;
[0057] S103 specifically includes:
[0058] The parameters of the individual decision-making layer are used as the input of BPNN, and the parameters of the group behavior layer are used as the output of BPNN to establish the functional mapping relationship between individual decision-making and group behavior.
[0059] The parameters of the group behavior layer are used as the input of BPNN, and the parameters of the top-level design layer are used as the output of BPNN to establish a functional mapping relationship between individual decision-making and group behavior.
[0060] S103 specifically includes:
[0061] Using the formula Determine the input-output relationship of BPNN;
[0062] Among them, x i Indicates the input value; y k Indicates expected value; Represents the weight between input layer i and hidden layer j; represents the weight between hidden layer j and output layer k; a j represents the threshold of hidden layer j; b k represents the threshold of output layer k, n represents the number of hidden layer nodes, m represents the number of input layer nodes, and k represents the number of output layer nodes.
[0063] S104, taking the task completion, design cost and design efficiency in the parameters of the top-level design layer as optimization targets, taking the parameter range of the individual decision layer as constraint conditions, and constructing the objective function of multi-objective optimization based on the functional mapping relationship between individual decision and group behavior and the functional mapping relationship between group behavior and the top-level design;
[0064] S104 specifically includes:
[0065] Using the formula Construct objective functions for multi-objective optimization;
[0066] Among them, Z1 represents the task completion, Z2 represents the design cost, Z3 represents the design efficiency, and p and q are constants.
[0067] The parameter range of the parameters X1, X2, ..., X5 of the individual decision layer is:
[0068]
[0069] S105, based on the second generation non-dominated sorting genetic algorithm NSGA-Ⅱ, a global search is performed on the objective function of multi-objective optimization to determine the global optimal solution. The core technologies of NSGA-Ⅱ include fast non-dominated sorting, crowding calculation and elite retention strategy, which can improve the diversity of the population and reduce the complexity of the optimization strategy.
[0070] To test this method, Figure 5 The NetLogo simulation platform shown in the figure obtains data, and Matlab2021a is used to perform modeling and optimization under Window 10. Tables 1-3 respectively show the normalized individual decision, group behavior, and top-level design parameters of the UAV cluster process behavior.
[0071] Table 1
[0072]
[0073]
[0074] Table 2
[0075]
[0076] Table 3
[0077]
[0078] In the process behavior modeling stage of drone clusters, the 100 groups of sample data in Table 1-3 are divided into the first 70 groups as training samples and the last 30 groups as test samples. The hidden layer transfer function of BPNN is Sigmoid function, and the output layer transfer function is Purelin function.
[0079] (1) “Individual decision-making-group behavior” mapping modeling: Individual decision parameters are used as the input of BPNN, and group behavior parameters are used as the output. The fitting effect between the actual value and the predicted value is shown in Figure 2. Figure 6 shown.
[0080] (2) “Group behavior-top-level design” mapping modeling: The group behavior parameters are used as the input of BPNN, and the top-level design parameters are used as the output. The fitting effect between the true value and the predicted value is shown in Figure 7 shown.
[0081] In the process behavior optimization stage of drone swarm, the established BPNN model is used as the mapping function in the multi-objective algorithm. The multi-objective optimization Pareto frontier is as follows: Figure 8 The results show that this method can effectively realize the behavior prediction of drone clusters, with an average prediction accuracy of 91%, and can explore the frontier of the Pareto solution set of the top-level optimization of drone clusters to avoid falling into the local optimal value, which provides a basis for further analysis of optimization results and adjustment of individual decision parameters.
[0082] Fig. 9 This is a schematic diagram of the structure of a UAV cluster process behavior modeling and collaborative optimization system provided by the present invention, such as Fig. 9 As shown, the present invention provides a UAV cluster process behavior modeling and collaborative optimization system, including:
[0083] The hierarchical decomposition module 901 is used to hierarchically decompose the drone swarm system using a top-down layer-by-layer decomposition strategy to determine the individual decision-making layer, the group behavior layer and the top-level design layer; the parameters of the individual decision-making layer include: attack capability, defense capability, mobility capability, communication capability and obstacle avoidance capability; the parameters of the group behavior layer include: group size, group density, clustering coefficient and average distance; the parameters of the top-level design layer include: task completion, design cost and design efficiency;
[0084] The sample data acquisition module 902 is used to acquire sample data of the drone cluster system using the NetLogo simulation platform; the sample data includes: parameters of the individual decision-making layer, parameters of the group behavior layer, and parameters of the top-level design layer;
[0085] A function mapping relationship establishment module 903 is used to establish a function mapping relationship between individual decision-making and group behavior and a function mapping relationship between group behavior and top-level design by using a back propagation neural network BPNN;
[0086] The objective function construction module 904 of multi-objective optimization is used to construct the objective function of multi-objective optimization based on the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design, taking the task completion, design cost and design efficiency in the parameters of the top-level design layer as optimization targets, taking the parameter range of the individual decision-making layer as constraint conditions, and constructing the objective function of multi-objective optimization based on the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design;
[0087] The global optimal solution determination module 905 is used to perform a global search on the objective function of multi-objective optimization based on the second generation non-dominated sorting genetic algorithm NSGA-Ⅱ to determine the global optimal solution.
[0088] The UAV swarm process behavior modeling and collaborative optimization system provided by the present invention also includes:
[0089] The data normalization module is used to perform normalization processing on the sample data.
[0090] The function mapping relationship establishing module 903 specifically includes:
[0091] The first function mapping relationship establishment unit is used to establish the function mapping relationship between individual decision-making and group behavior by taking the parameters of the individual decision-making layer as the input of BPNN and taking the parameters of the group behavior layer as the output of BPNN;
[0092] The second function mapping relationship establishing unit is used to establish the function mapping relationship between individual decision-making and group behavior by taking the parameters of the group behavior layer as the input of BPNN and taking the parameters of the top design layer as the output of BPNN.
[0093] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0094] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of the present invention. At the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.
Claims
1. A method for modeling and collaborative optimization of UAV swarm process behavior, characterized in that: include: The drone swarm system is decomposed layer by layer from top to bottom to determine the individual decision-making layer, group behavior layer and top-level design layer; personal The parameters of the decision-making layer include: attack capability, defense capability, mobility, communication capability and obstacle avoidance capability; the parameters of the group behavior layer include: group size, group density, clustering coefficient and average distance; the parameters of the top-level design layer include: task completion, design cost and design efficiency; The NetLogo simulation platform is used to obtain sample data of the drone cluster system; the sample data includes: parameters of the individual decision-making layer, parameters of the group behavior layer, and parameters of the top-level design layer; The back propagation neural network (BPNN) is used to establish the functional mapping relationship between individual decision-making and group behavior, and the functional mapping relationship between group behavior and top-level design. Taking the task completion, design cost and design efficiency in the parameters of the top-level design layer as the optimization goals, and the parameter range of the individual decision layer as the constraint condition, the objective function of multi-objective optimization is constructed based on the functional mapping relationship between individual decision-making and group behavior and the functional mapping relationship between group behavior and top-level design; Based on the second generation non-dominated sorting genetic algorithm NSGA-Ⅱ, a global search is performed on the objective function of multi-objective optimization to determine the global optimal solution.
2. The method for modeling and collaborative optimization of UAV swarm process behavior according to claim 1 is characterized in that: The method of using the NetLogo simulation platform to obtain sample data of the drone cluster system also includes: The sample data is normalized.
3. The method for modeling and collaborative optimization of UAV swarm process behavior according to claim 1, characterized in that: The back propagation neural network BPNN is used to respectively establish the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design, specifically including: The parameters of the individual decision-making layer are used as the input of BPNN, and the parameters of the group behavior layer are used as the output of BPNN to establish the functional mapping relationship between individual decision-making and group behavior. The parameters of the group behavior layer are used as the input of BPNN, and the parameters of the top-level design layer are used as the output of BPNN, and a functional mapping relationship between group behavior and top-level design is established.
4. The method for modeling and collaborative optimization of UAV swarm process behavior according to claim 3 is characterized in that: The back propagation neural network BPNN is used to respectively establish the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design, specifically including: Using the formula Determine the input-output relationship of BPNN; Among them, x i Indicates the input value; y k Indicates expected value; Represents the weight between input layer i and hidden layer j; represents the weight between hidden layer j and output layer k; a j represents the threshold of hidden layer j; b k represents the threshold of output layer k, n represents the number of hidden layer nodes, m represents the number of input layer nodes, and k represents the number of output layer nodes.
5. The method for modeling and collaborative optimization of UAV swarm process behavior according to claim 3 is characterized in that: The objective function of multi-objective optimization is constructed based on the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design, taking the task completion, design cost and design efficiency in the parameters of the top-level design layer as the optimization target, taking the parameter range of the individual decision-making layer as the constraint condition, and constructing the objective function of multi-objective optimization based on the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design, specifically including: Using the formula Construct objective functions for multi-objective optimization; Among them, Z1 represents the task completion, Z2 represents the design cost, Z3 represents the design efficiency, and p and q are constants.
6. A UAV swarm process behavior modeling and collaborative optimization system, characterized in that: include: The hierarchical decomposition module is used to decompose the drone cluster system hierarchically using a top-down layer-by-layer decomposition strategy to determine the individual decision-making layer, group behavior layer, and top-level design layer; The parameters of the individual decision-making layer include: attack capability, defense capability, mobility, communication capability, and obstacle avoidance capability; the parameters of the group behavior layer include: group size, group density, clustering coefficient, and average distance; the parameters of the top-level design layer include: task completion, design cost, and design efficiency; The sample data acquisition module is used to obtain sample data of the drone cluster system using the NetLogo simulation platform; the sample data includes: parameters of the individual decision-making layer, parameters of the group behavior layer, and parameters of the top-level design layer; A function mapping relationship establishment module is used to establish the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design by using a back propagation neural network (BPNN); The objective function construction module of multi-objective optimization is used to construct the objective function of multi-objective optimization based on the function mapping relationship between individual decision-making and group behavior and the function mapping relationship between group behavior and top-level design, taking the task completion, design cost and design efficiency in the parameters of the top-level design layer as the optimization objectives and the parameter range of the individual decision-making layer as the constraint condition; The global optimal solution determination module is used to perform a global search for the objective function of multi-objective optimization based on the second-generation non-dominated sorting genetic algorithm NSGA-Ⅱ to determine the global optimal solution.
7. The UAV swarm process behavior modeling and collaborative optimization system according to claim 6 is characterized in that: Also includes: The data normalization module is used to perform normalization processing on the sample data.
8. The UAV swarm process behavior modeling and collaborative optimization system according to claim 6 is characterized in that: The function mapping relationship establishment module specifically includes: The first function mapping relationship establishment unit is used to establish the function mapping relationship between individual decision-making and group behavior by taking the parameters of the individual decision-making layer as the input of BPNN and taking the parameters of the group behavior layer as the output of BPNN; The second function mapping relationship establishing unit is used to establish the function mapping relationship between the group behavior and the top-level design by taking the parameters of the group behavior layer as the input of the BPNN and the parameters of the top-level design layer as the output of the BPNN.
