摘要
离散多智能体系统的一致性问题是复杂动态大系统中重要问题,对收敛速度更是研究的重点。设计了一种适用于任意无向连通结构的多智能体一致性预测控制算法μ-MPC,提高智能体的一致性收敛速度并扩展了离散系统采样步长的取值范围,增强了系统的稳定性。在一定理论分析的基础上,通过大量的仿真实例和与经典模型的比较,研究了算法在固定拓扑和切换拓扑系统中的快速性和鲁棒性。特别对于全连通结构系统有一定的参考价值。
Consensus of distributed multi-agent system is a significant problem in the complex dynamics system,and the speed of consensus is a focused issue.A Model Predictive Control consensus algorithm μ-MPC is designed for any undirected graph with strongly connected topology.This algorithm can speed up the consensus extremely and enlarge the range of step size,which improves the stability of the system.Based on some theoretical analysis,a large amount of simulation examples demonstrate the fastness,robustness of μ-MPC compared with other protocols under fixed topology as well as switching systems.Particularly,under this algorithm,the system with complete graph can reach consensus in one step.
出处
《计算机仿真》
CSCD
北大核心
2010年第12期186-190,210,共6页
Computer Simulation
基金
国家自然科学基金(60674041
60934007)
学校博士学科点专项科研基金(20070248004)
关键词
多智能体一致性
预测控制
仿真
快速性
切换拓扑
Consensus of multi-agent system
MPC
Simulation
Fastness
Switching topology
