Discrete heterogeneous multi-agent system grouping consistency control method based on cooperation-competition relationship

Abstract

The invention discloses a discrete heterogeneous multi-agent system grouping consistency control method based on a cooperation-competition relation. Belongs to the technical field of artificial intelligence control. The method comprises the following steps: 1, establishing a heterogeneous multi-agent system model composed of m second-order agents and n first-order agents; 2, establishing a communication topology of the heterogeneous multi-agent system; 3, establishing a cooperation-competition relation functional expression between the intelligent agents; 4, establishing a heterogeneous multi-agent system consistency controller based on a competition-cooperation relationship; and 5, applying the established controller to the established heterogeneous multi-agent system model to realize the consistency control of the heterogeneous multi-agent system based on the competition-cooperation relationship. According to the method, communication between heterogeneous multi-agent systems is carried out based on the competition-cooperation relationship between the agents, and under the condition that packet loss and the interaction strength between the agents are considered, sufficient conditions for ensuring the agent grouping consistency are obtained.

Description

Technical Field

[0001] The present invention belongs to the technical field of multi-agent system control, and particularly relates to a grouping consensus control method for a discrete heterogeneous multi-agent system based on a cooperation-competition relationship. Background Art

[0002] A multi-agent system is an aggregate composed of a certain number of agents. Each individual independent agent has the ability to cooperate with other agents while having a certain degree of autonomy. Therefore, the entire agent system can jointly complete a specified task or achieve a common goal through mutual communication. Since a multi-agent system can complete complex tasks through mutual interaction and collective decision-making, it has a wide range of applications in many fields, such as unmanned aerial vehicle formation control, distributed sensor networks, satellite formation control, robot formation control, etc.

[0003] As a basic problem in the field of multi-agent control, multi-agent consensus control has received extensive attention. Multi-agent consensus control means that agents reach and maintain a consistent state during the process of task execution through communication and information sharing. When a multi-agent system is working, two types of time delays will occur. The first is communication delay, that is, the time delay generated when agents exchange information with each other. The second is input delay, that is, when an agent receives an input signal, the time delay generated due to the influence of external disturbances. Both of the above time delays will affect the system stability and coordination. Therefore, when studying grouping consensus, communication delay and input delay are the main parameters that need to be considered.

[0004] Currently, many studies on the grouping consensus of multi-agent systems are aimed at homogeneous multi-agent systems. That is, all information transmission and exchange are restricted to be carried out within the same subgroup. Obviously, this is unrealistic. In practical applications, each agent has its own dynamic characteristics, and each agent will be affected by external environment disturbances, resulting in no two exactly the same multi-agent systems. Moreover, in order to better coordinate and reach a common goal, agents in different subgroups also need to exchange information. Therefore, heterogeneous multi-agent systems have a wider range of application and more profound practical significance.

[0005] The cooperation relationship and competition relationship between agents often coexist. In an agent system based on a cooperation-competition relationship, agents within the same group cooperate and approach each other, while agents in different groups compete and move away from each other. Since the grouping consensus based on the cooperation-competition relationship is more consistent with the actual behavior of objects, the research on it is more meaningful. Summary of the Invention

[0006] To solve the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a grouping consensus control method for a discrete heterogeneous multi-agent system based on cooperation-competition relationships, and sufficient conditions for achieving grouping consensus are obtained under the condition that the vertex in-degree balance is not satisfied.

[0007] To achieve the above purpose, the present invention adopts the following technical solutions:

[0008] A grouping consensus control method for a discrete heterogeneous multi-agent system based on cooperation-competition relationships includes the following steps:

[0009] Step 1, establish a heterogeneous multi-agent system model composed of nP first-order agents and m second-order agents. The heterogeneous multi-agent system model is as follows:

[0010] x i (k + 1) = x i (k) + u i (k), i ∈ σ2

[0011] where σ1 = {1, 2,..., m}, σ2 = {m + 1, m + 2,..., m + n}. σ = σ1 ∪ σ2, σ1 ∩ σ2 = φ. x i (k) ∈ R N , v i (k) ∈ R N , u i (k) ∈ R N are respectively the state, speed, and input control of the i-th agent.

[0012] Step 2, establish the communication topology of the heterogeneous multi-agent system: It is described by a weighted topology graph. Let G = (V(G), E(G), A) be a weighted topology graph with n vertices, used to describe the information interaction between agents in the multi-agent system. The vertex set V(G) = {v1, v2,..., v m , v m+1 ,..., v m+n}, representing m + n agents. The edge set E(G) ∈ V × V. If e ij = (v i , v j ) ∈ E(G), it means that agent i can receive the transmission information of agent j, otherwise A = (a ij ) ∈ R (m+n)×(m+n) is the adjacency matrix of G. In particular, there is no self-loop in this graph. If a ij > 0, if and only if e ij ∈ E(G). N i = {v j ∈ v|eij ∈ E(G)}, respectively represent the neighbor set of node i and the in-degree of vertex v i of, D = diag{d1, d2, …, d m , d m+1 , …, d m+n} is the in-degree matrix.

[0013] Step 3, establish a consensus controller for the heterogeneous multi-agent system based on the competition - cooperation relationship, which is expressed as:

[0014]

[0015] Step 4, apply the controller established in Step 3 to the heterogeneous multi-agent system model established in Step 1 to achieve the consensus control of the heterogeneous multi-agent system based on the competition - cooperation relationship.

[0016] Since the present invention adopts a discrete heterogeneous multi-agent system grouping consensus control method based on the cooperation - competition relationship, the following beneficial effects compared with the prior art can be obtained:

[0017] By considering the time delay, the present invention obtains a sufficient condition for the heterogeneous multi-agent system based on the competition - cooperation relationship to achieve grouping consensus, which promotes the full play of the characteristics and functions of the multi-agent system while achieving consensus. Description of the Drawings

[0018] Figure 1 is a schematic flow chart of a discrete heterogeneous multi-agent system grouping consensus control method based on the cooperation - competition relationship provided by an embodiment of the present invention;

[0019] Figure 2 is a schematic communication topology diagram of the heterogeneous multi-agent system provided by an embodiment of the present invention;

[0020] Figure 3 is a schematic position diagram of the heterogeneous multi-agent system provided by an embodiment of the present invention;

[0021] Figure 4 is a schematic speed diagram of the heterogeneous multi-agent system provided by an embodiment of the present invention; Detailed Embodiment

[0022] In order to better enable those skilled in the field of multi-agent control to understand the method of the present invention, the present invention will be described in more detail below through invention embodiments and drawings. It should be emphasized that the embodiments of the present invention are carried out on the premise of the technical solution of the present invention, and are only used to illustrate the present invention, rather than limiting its scope. The embodiments of the present invention are not limited thereto.

[0023] As shown in Figure 1 the figure, a method for grouping consistency control of a discrete heterogeneous multi-agent system based on a cooperation-competition relationship according to the present invention includes the following steps:

[0024] Step 1: Establish a heterogeneous multi-agent system model composed of n first-order agents and m second-order agents. The heterogeneous multi-agent system model is as follows: x i (k + 1) = x i (k) + u i (k), i ∈ σ2

[0025] where σ1 = {1, 2,..., m}, σ2 = {m + 1, m + 2,..., m + n}. σ = σ1 ∪ σ2, σ1 ∩ σ2 = φ. x i (k) ∈ R N , v i (k) ∈ R N , u i (k) ∈ R N are respectively the state, speed, and input control of the i-th agent.

[0026] Step 2: Establish the communication topology of the heterogeneous multi-agent system: It is described by a weighted topology graph. Let G = (V(G), E(G), A) be a weighted topology graph with n vertices, used to describe the information interaction between agents in the multi-agent system. The vertex set V(G) = {v1, v2,..., v m , v m+1 ,..., v m+n}, representing m + n agents. The edge set E(G) ∈ V × V. If e ij = (v i , v j ) ∈ E(G), it means that agent i can receive the transmission information of agent j. Otherwise A = (a ij ) ∈ R (m+n)×(m+n) is the adjacency matrix of G. In particular, there is no self-loop in this graph. If a ij > 0, if and only if e ij ∈ E(G). N i = {v j ∈ v | e ij ∈ E(G)}, respectively represent the neighbor set of node i and the in-degree of vertex v i . D = diag{d1, d2,..., d m , d m+1 ,..., d m+n} is the in-degree matrix.

[0027] Step 3: Establish the competition - cooperation relationship between agents. Use x j (k) - x i (k) to represent that agent i and agent j are in a cooperative relationship, and use x j (k) + x i (k) to represent that agent i and agent j are in a competitive relationship, and the same meaning applies to the speed.

[0028] Step 4: Establish a consensus controller for heterogeneous multi - agent systems based on the competition - cooperation relationship:

[0029]

[0030] Step 5: Apply the controller established in Step 4 to the heterogeneous multi - agent system model established in Step 1 to achieve the consensus control of the heterogeneous multi - agent system based on the competition - cooperation relationship.

[0031] The following is a simulation example of the method provided in this instance. The simulation example considers that the system has a total of 2 groups, G1 and G2, and a total of 8 agents. The 8 agents belong to G1 and G2 respectively, as shown specifically in Figure 2 . Among them, i = 1, 2, 3, 4 are second - order agents. i = 5, 6, 7, 8 are first - order agents. Each part consists of first - order and second - order agents.

[0032] Consider the case of α≠β and γ≠φ. Select p = 0.5, ε ij = 1, {α1, α2, α3, α4, α5, α6} = {0.9, 0.2, 0.25, 0.6, 2.9, 0.4}, {β1, β2, β3, β4, β5, β6} = {2, 0.8, 1.7, 0.8, 3.5, 0.6} {γ1, γ2, γ3, γ4, γ5, γ6} = {0.35, 0.8, 0.2, 1.5, 1.9, 2.3} {φ1, φ2, φ3, φ4, φ5, φ6} = {1, 1.2, 0.8, 2, 2.4, 3.3}

[0033] By observing Figure 3 and Figure 4 the curves in, it can be clearly seen the effect of the consensus control method for heterogeneous multi - agent systems based on the competition - cooperation relationship in the embodiments of the present invention. Through this control method, the heterogeneous multi - agent system composed of first - order and second - order agents can achieve group consensus and optimize the control effect.

[0034] The technical solution of the present invention is not limited to the previously disclosed technical means, but includes technical solutions composed of various combinations of these technical features. The preferred embodiments of the present invention have been introduced in detail above. However, it is worth emphasizing that those of ordinary skill in the relevant technical field may still make various improvements and optimizations without departing from the principle of the present invention. These improvements and optimizations are also regarded as part of the protection scope of the present invention. Therefore, the protection scope of this invention patent should cover all equivalent embodiments that conform to the combination of technical features, rather than being limited to the specifically disclosed embodiments.

Claims

1. A group consistency control method for discrete heterogeneous multi-agent systems based on cooperation-competition relationship, characterized by: The steps include: Step 1: Establish a heterogeneous multi-agent system model consisting of n first-order agents and m second-order agents. The dynamics model of the heterogeneous multi-agent system is as follows: x i (k+1)=x i (k)+u i (k),i∈σ2 Among them, σ1={1,2,…,m}, σ2={m+1,m+2,…,m+n}. σ=σ1∪σ2, σ1∩σ2=φ. x i (k)∈R N , v i (k)∈R N ,u i (k)∈R N are the state, speed and input control of the ith agent respectively. Step 2: Establish the communication topology of the heterogeneous multi-agent system: The communication topology can be described by a weighted topology graph. Let G = (V(G), E(G), A) be a weighted topology graph with n vertices, which is used to describe the information interaction between agents in the multi-agent system. The vertex set V(G) = {v1, v2, …, v m ,v m+1 ,…,v m+n }, representing m+n agents. The edge set E(G)∈V×V, if e ij =(v i ,v j )∈E(G), which means that agent i can receive the transmission information of agent j, otherwise A=(a ij )∈R (m+n)×(m+n) is the adjacency matrix of G. In particular, there is no self-loop in this graph. ij >0, if and only if e ij ∈E(G). N i = {v j ∈v|e ij ∈E(G)}, Represent the neighbor set of node i and vertex v respectively i The in-degree, D = diag{d1,d2,…,d m ,d m+1 ,…,d m+n } is the in-degree matrix. L = DA is the Laplacian matrix of graph G. Step 3: Establish the cooperation-competition relationship between agents. j (k)-x i (k) indicates that agent i and agent j are in a cooperative relationship, and x j (k)+x i (k) indicates that agent i and agent j are in a competitive relationship, which has the same meaning for speed. Step 4: Establish a heterogeneous multi-agent system consistency controller based on competition-cooperation relationship: where w i (k) is the designed adaptive controller. i >0,β i >0,γ i >0,φ i >0 is a variable control parameter. ij represents the intensity of cooperation or competition between agents i and j. p∈(0,1] is the packet loss rate, s i It represents the set of neighbors that belong to the same group as agent i. All agents in this set have a cooperative relationship with each other. i It represents the set of neighbors that belong to different groups from agent i. All agents in this set have a competitive relationship with each other. Step five, apply the established controller to the established heterogeneous multi-agent system model to achieve consistency control of the heterogeneous multi-agent system based on competition-cooperation relationship.

Images