A research method for decentralized consistency of heterogeneous multi-agent systems based on complex relationships

Abstract

The present invention belongs to the technical field of intelligent control research, and specifically relates to a decentralized research method for consistency of heterogeneous multi-agent systems based on complex relationships, comprising the following steps: Step S100: Analyze and study the random factors in the multi-agent system and the complex relationship characteristics between agents to obtain a dynamic model of the multi-agent system; Step S200: Based on the dynamic model of the multi-agent system, obtain sufficient conditions for the consistency of the multi-agent system; Step S300: Analyze the decentralized behavior of the multi-agent system and establish a decentralized model for the consistency of the multi-agent system; The present invention establishes an agent group behavior factor model by considering the random factors in the multi-agent system and the complex relationship characteristics between agents, clarifies the influence of the agent's own dynamics, network topology, and communication mechanism on the design of decentralized control protocols in the consistency research, and improves the theoretical basis and implementation method of system performance.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent control research, and specifically relates to a decentralized research method for consistency of heterogeneous multi-agent systems based on complex relationships. Background Art

[0002] Multi-agent system is a typical form of complex system, and consistency is one of the main dynamic behaviors of complex system. How to realize the decentralized strategy of consistency of heterogeneous multi-agent system under the background of random factors and complex relationships is one of the key issues in realizing the practical application of multi-agent system. This invention is based on the existing research framework, analyzes the characteristics of random factors, complex relationships, deeply understands the behavior mechanism of agents, establishes a dynamic evolution model, and explores the role and influence of decentralization on the consistency of multi-agent system under random factors and complex relationships from the perspective of theory, simulation and application, and further answers the optimization methods and strategies of feedback complex systems. Summary of the invention

[0003] The purpose of the present invention is to provide a decentralized research method for consistency of heterogeneous multi-agent systems based on complex relationships, so as to solve the problems existing in the background technology.

[0004] In order to achieve the above technical objectives, the technical solution adopted by the present invention is as follows:

[0005] A decentralized research method for consistency of heterogeneous multi-agent systems based on complex relationships includes the following steps:

[0006] Step S100: Analyze and study the random factors in the multi-agent system and the complex relationship characteristics between agents to obtain a dynamic model of the multi-agent system;

[0007] Step S200: Based on the multi-agent system dynamics model, sufficient conditions for the consistency of the multi-agent system are obtained;

[0008] Step S300: Analyze the decentralized behavior of the multi-agent system and establish a decentralized model of the consistency of the multi-agent system.

[0009] The “random factors in the multi-agent system” in step S100 include, but are not limited to, network congestion, packet loss, random interference, malicious attacks, delays, and communication interruptions;

[0010] The "complex relationship between intelligent agents" in step S100 includes, but is not limited to, factors such as cooperation-competition, heterogeneous systems, their own dynamic changes, and changes in network topology.

[0011] The step S100 of "obtaining a multi-agent system dynamics model" mainly includes the following sub-steps:

[0012] Step S110: Determine the random factors that affect the multi-agents and the types of complex relationships between the multi-agents, and derive a dynamics model that reflects the system of agents based on their dynamic relationships;

[0013] Step S120: using simulation software to verify the correctness of the established dynamic model.

[0014] The step S110 includes the following sub-steps:

[0015] Step S111: describing the dynamic behavior of a multi-agent system including multiple agents;

[0016] Step S112: designing a group consistency control protocol for a multi-agent system;

[0017] Step S113: Model validity verification.

[0018] The step S111 is specifically as follows:

[0019] Assume that the heterogeneous multi-agent system contains m+n agents, where m nodes are second-order nodes and n nodes are first-order nodes. Its dynamic behavior is described as follows:

[0020] Formula 1:

[0021]

[0022] Among them: σ1={1}σ1={1,2,...,n}, σ2={n+1,n+2,...,n+m}, and σ=σ1∪σ2;

[0023] x i (k)∈R N , v i (k)∈R N ,u i (k)∈R N Respectively represent the position state, velocity, and control input of node i;

[0024] n first-order nodes belong to group σ2, m second-order nodes belong to group σ1, the nodes within the group are in a cooperative relationship and have the same goals, the nodes between the groups are in a competitive relationship, and the relationship between the two groups is negatively correlated.

[0025] Step S112 is specifically as follows:

[0026] Based on the cooperation-competition relationship, considering the complex environment faced by the multi-agent system, a group consistency control protocol is designed, as follows:

[0027]

[0028] In step S120, an adjacency matrix for coupling between agents is also set to verify the validity of the model. The adjacency matrix for coupling between agents is set as follows:

[0029]

[0030] The step S200 also includes the following sub-steps:

[0031] Step S210: using graph algebra to represent the communication between the agents and determine the cluster model under the complex relationship between different individuals of the agents;

[0032] Step S220: Using stability theory, determine whether the individuals in the cluster converge to a bounded region within a finite time.

[0033] The present invention establishes an agent group behavior factor model by taking into account the random factors in the multi-agent system and the complex relationship characteristics between agents, clarifies the influence of the agent's own dynamics, network topology, and communication mechanism on the design of distributed control protocols in consistency research, and improves the theoretical basis and implementation method of system performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present invention is further illustrated by means of the non-limiting examples given in the accompanying drawings.

[0035] Figure 1 It is a schematic diagram of the process of the present invention;

[0036] Figure 2 This is a schematic diagram of the grouping consistency of the present invention. Figure 1 ;

[0037] Figure 3 This is a schematic diagram of the grouping consistency of the present invention. Figure 2 ;

[0038] Figure 4 This is a schematic diagram of the grouping consistency of the present invention. Figure 3 ; DETAILED DESCRIPTION

[0039] In order to enable those skilled in the art to better understand the present invention, the technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0040] A decentralized research method for consistency of heterogeneous multi-agent systems based on complex relationships includes the following steps:

[0041] Step S100: Analyze and study the random factors in the multi-agent system and the complex relationship characteristics between agents to obtain a dynamic model of the multi-agent system;

[0042] The “random factors in the multi-agent system” in step S100 include, but are not limited to, network congestion, packet loss, random interference, malicious attacks, delays, and communication interruptions;

[0043] The "complex relationship between agents" in step S100 includes, but is not limited to, factors such as cooperation-competition, heterogeneous systems, their own dynamic changes, and network topology changes;

[0044] The step S100 of "obtaining a multi-agent system dynamics model" mainly includes the following sub-steps:

[0045] Step S110: Determine the random factors that affect the multi-agents and the types of complex relationships between the multi-agents, and derive a dynamics model that reflects the agent system based on their dynamic relationships, as follows:

[0046] Step S111: Describing the dynamic behavior of a multi-agent system including multiple agents, as follows:

[0047] Assume that the heterogeneous multi-agent system contains m+n agents, where m nodes are second-order nodes and n nodes are first-order nodes. Its dynamic behavior is described as follows:

[0048] Formula 1:

[0049]

[0050] Among them: σ1={1}σ1={1,2,...,n}, σ2={n+1,n+2,...,n+m}, and σ=σ1∪σ2;

[0051] x i (k)∈R N , v i (k)∈R N ,u i (k)∈R N Respectively represent the position state, velocity, and control input of node i;

[0052] n first-order nodes belong to group σ2, m second-order nodes belong to group σ1, the nodes within the group are in a cooperative relationship, they have the same goal, the nodes between the groups are in a competitive relationship, and the relationship between the two groups is negatively correlated;

[0053] Step S112: Design a group consistency control protocol for the multi-agent system, as follows:

[0054] Based on the cooperation-competition relationship and considering the complex environment faced by the multi-agent system, the group consistency control protocol is designed as follows:

[0055]

[0056] Through the application of the adjacency matrix, the state trajectory of the system is as follows Figure 2 As shown, the system achieves group consistency;

[0057] Step S120: Use simulation software to verify the correctness of the established dynamic model. In step S120, an adjacency matrix for coupling between agents is set to verify the validity of the model. The adjacency matrix for coupling between agents is set as follows:

[0058]

[0059] Through the set adjacency matrix, the state trajectory of the system is obtained as follows Figures 2 to 4 As shown, it can be seen that the system achieves group consistency.

[0060] Step S200: Based on the multi-agent system dynamics model, sufficient conditions for the consistency of the multi-agent system are obtained, which is specifically performed using the following sub-steps:

[0061] Step S210: using graph algebra to represent the communication between the agents and determine the cluster model under the complex relationship between different individuals of the agents;

[0062] Step S220: using stability theory, determining that individuals in the cluster converge to a bounded region within a finite time;

[0063] Step S300: Analyze the decentralized behavior of the multi-agent system and establish a decentralized model of the consistency of the multi-agent system.

[0064] The present invention has the following advantages:

[0065] ① Considering the influence of multiple factors such as network congestion, packet loss, random interference, malicious attacks, delay, cooperation and competition, heterogeneous systems, dynamic changes of the system, and changes in network topology, a more realistic multi-agent system dynamics model is established;

[0066] ② It proposes to establish a model of agent group behavior factors, clarifies the influence of the agent's own dynamics, network topology, and communication mechanism on the design of decentralized control protocols in consistency research, and provides the theoretical basis and implementation methods for improving system performance;

[0067] ③ Use graph algebra, stability theory and decentralized control theory to find sufficient conditions for achieving consistency of multi-agent systems. Use decentralized control theory to study the decentralized mechanism of multi-agent consistency and further optimize the stability of multi-agent systems.

[0068] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A decentralized research method for consistency of heterogeneous multi-agent systems based on complex relationships, characterized by: The following steps are involved: Step S100: Analyze and study the random factors in the multi-agent system and the complex relationship characteristics between agents to obtain a dynamic model of the multi-agent system; The "random factors in the multi-agent system" in step S100 include network congestion, packet loss, random interference, malicious attacks, delays, and communication interruptions; The "complex relationship between agents" in step S100 includes cooperation-competition, heterogeneous systems, dynamic changes of the agents themselves, and changes in network topology; The step S100 of "obtaining a multi-agent system dynamics model" mainly includes the following sub-steps: Step S110: Determine the random factors that affect the multi-agents and the types of complex relationships between the multi-agents, and derive a dynamics model that reflects the system of agents based on their dynamic relationships; The step S110 includes the following sub-steps: Step S111: Describing the dynamic behavior of a multi-agent system including multiple agents, specifically: Assume that the heterogeneous multi-agent system contains m+n agents, where m nodes are second-order nodes and n nodes are first-order nodes. Its dynamic behavior is described as follows: Among them: σ1={1,2,...,n}, σ2={n+1,n+2,...,n+m}, and σ=σ1∪σ2; x i (k)∈R N , v i (k)∈R N ,u i (k)∈R N Respectively represent the position state, velocity, and control input of node i; n first-order nodes belong to group σ2, m second-order nodes belong to group σ1, the nodes within the group are in a cooperative relationship, they have the same goal, the nodes between the groups are in a competitive relationship, and the relationship between the two groups is negatively correlated; Step S112: Design a group consistency control protocol for the multi-agent system, specifically: Based on the cooperation-competition relationship and considering the complex environment faced by the multi-agent system, the group consistency control protocol is designed as follows: Step S120: using simulation software to verify the correctness of the established dynamic model; Step S200: Based on the multi-agent system dynamics model, sufficient conditions for the consistency of the multi-agent system are obtained; Step S300: Analyze the decentralized behavior of the multi-agent system and establish a decentralized model of the consistency of the multi-agent system.

2. According to claim 1, a method for decentralized research on consistency of heterogeneous multi-agent systems based on complex relationships is characterized by: In step S120, an adjacency matrix for coupling between agents is set to verify the validity of the model. The adjacency matrix for coupling between agents is set as follows:

3. According to claim 2, a method for decentralized research on consistency of heterogeneous multi-agent systems based on complex relationships is characterized by: The step S200 also includes the following sub-steps: Step S210: using graph algebra to represent the communication between the agents and determine the cluster model under the complex relationship between different individuals of the agents; Step S220: Using stability theory, determine whether the individuals in the cluster converge to a bounded region within a finite time.

Images