合作-竞争异质多智能体系统分组一致协同演化研究

Group consensus of multi-agent systems is one of the major dynamic behaviors of complex systems, and is also a prerequisite and fundamental issue for distributed coordinated control of complex systems. Based on the complex relationship among the agents and agents' heterogeneity, how to achieve group consensus under the situation of limited and discontinuous communication is one of the key problems in the practical applications of multi-agent systems. In this project, the advantage of impulsive control and adaptive pinning control theories will be effectively applied to investigate the group consensus for a class of heterogeneous multi-agent systems. Based on the cooperative and competitive relationships among the agents, the emergence mechanism of swarm intelligence and dynamic model of multi-agent systems will be established which can reveal the internal evolution of collective behaviors and represent the interaction between the agents and the groups as well. The criteria which guarantee the achievement of group consensus will be addressed under complex environment, such as time delays, random interferences and time-varying topologies. From the obtained criteria, the influence of these factors on the group consensus can also be clarified. Furthermore, the optimization method and strategy of coordinated control about the convergence and convergence rate of multi-agent systems will be investigated at the same time. The research results of the project will enrich and develop new theories and technologies for the coordinated control of multi-agent systems, and provide new ideas and methods for the theoretical analysis, design and control of complex systems.

多智能体系统的分组一致性是复杂系统的主要动力学行为之一，也是复杂系统分布式协调控制的前提与根本性问题。如何在通信受限、非连续通信背景下，基于智能体间的复杂关系与异质特性，实现多智能体系统的分组一致性，是多智能体系统实际应用的关键问题之一。本申请项目拟充分发挥脉冲控制与自适应、牵制控制的优势，基于智能体间的合作-竞争关系，深层次揭示群集现象的内在演化规律，建立群体智能涌现机制与动力学演化模型，客观体现智能体及分组间的通信交互，探索异质多智能体系统分组一致动力学演化规律与协作机制，获得系统在复杂网络环境下达到分组一致的条件判据。弄清楚系统拓扑结构、各类时延及随机干扰等因素对多智能体系统分组一致性的影响，探讨复杂系统分组一致的收敛性、收敛速度等协调控制的优化方法与策略。相关研究成果将丰富和发展多智能体系统协调控制的新理论、新技术，为复杂系统的设计、理论分析与控制提供新的思想与方法。

多智能体系统的分组一致性是复杂系统的主要动力学行为之一，也是复杂系统分布式协同控制的前提与根本性问题。如何在通信受限、非连续通信背景下，基于智能体间的复杂关系与异质特性，实现多智能体系统的分组一致性协同控制，是多智能体系统实际应用的关键问题之一。围绕项目的主要研究内容与拟解决的科学问题，在异质多智能体系统的协同、安全以及最优控制等方面开展了理论研究及应用方面的相关工作。在复杂系统建模及稳定性理论分析、安全控制策略与最优化控制算法设计等方面取得了一系列的研究成果。发挥脉冲控制、事件驱动控制与自适应、牵制控制的优势，基于智能体间的合作-竞争关系，深层次揭示群集现象的内在演化规律，建立群体智能涌现机制与动力学演化模型，客观体现智能体及分组间的通信交互，探索异质多智能体系统分组一致动力学演化规律与协作机制，获得系统在复杂网络环境下达到分组一致的条件判据。从系统复杂性、拓扑复杂性以及连接复杂性三个维度，解析了智能体间的耦合关系、通信时延、输入时延、随机干扰、拓扑结构等对复杂系统分组一致的影响关系，进一步弄清楚了复杂系统分组一致动力学演化规律与协作机制。本项目共发表学术论文36篇，申请专利22项，出版学术专著2部，相关研究成果将丰富和发展多智能体系统协调控制的新理论、新技术，为复杂系统的设计、理论分析与控制提供了新的思想与方法。