Research On Multi-agent Consensus Protocol Based On Neighborhood Information Selection

The multi-agent system has been the main research object in many subject areas.The research on multi-agent cooperative control theory based on switching topology is an important part of multi-agent system research and has been a hot topic in recent years.The consistency problem is the basis of cooperation and coordination control between agents.Therefore,the research on the coherence protocol designed for the consistency problem has important practical significance and theoretical value.The traditional coherence protocol based on switching topology requires the agent to refer to the state information of all the neighboring members to carry out the state evolution at the next moment.Therefore,the burden of computing and storage of the system is large,but the researchers have discovered that more agents are between the agents.The connection does not always promote faster convergence of the entire system,so designing a consensus protocol with a neighborhood information selection strategy can not only guarantee the same convergence rate but also reduce the energy consumption of the entire system.At the same time,whether the final convergence of a multi-agent system is closely related to the composition of the network topology diagram.Multi-agent system in different topology,the choice of neighborhood information will affect the performance of the coherence protocol,so the coherence protocol with different neighborhood information selection strategy has its counterpart to better use the coherence agreement.The topological structure of the performance.Firstly,based on the density of social centers,a class of coherence protocols with neighboring information selection strategies is designed,including three neighborhood information selection strategies.Strategy 1 only refers to the state of the agents in the neighborhood with the lowest density of social networks in the neighborhood.Based on strategy 1,strategy 2 further integrates the relative Euclidean distance between agents,and strategy 3 regards the ratio of the number of neighbors shared between the agent and each neighbor agent and the number of agents within the agent's neighbor as the weights and the social network.Density combination.Then this type of coherence protocol is simulated in two topological structures: random topology and uniform topology.The experimental results show that the performance of coherence protocols using this neighborhood information selection strategy varies significantly based on different topology structures.Then,a kind of coherence protocol based on the longest distance selection neighbor strategy is designed,including two kinds of neighborhood information selection strategies.The strategy 4 selects the farthest distance in each direction in the neighborhood from the perspective of all perceived directions of the agent.Neighbor agents participate in the evolution of the state.Strategy 5 divides all directions into multiple angle ranges,and chooses the neighbor with the longest distance to participate in the evolution from each angle range.Then this type of coherence protocol is simulated in two topological structures: random topology and uniform topology.The experimental results show that the performance of the coherence protocol based on different topological structures and choosing the neighbor strategy based on the longest distance is obviously different.Finally,two kinds of topological structure experiment results are summarized for two kinds of coherence protocols with neighborhood information selection strategy.