项目名称: 复杂多自主体系统的分析、优化与调控
项目编号: No.61273221
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 刘志新
作者单位: 中国科学院数学与系统科学研究院
项目金额: 80万元
中文摘要: 复杂系统研究被称为"二十一世纪的科学",而多自主体系统是复杂系统中备受瞩目的研究方向之一。多自主体系统广泛存在于生物、物理、工程及社会经济系统中。本项目将开展三方面的研究:一是多自主体系统集体行为的分析。我们将从三类基本而又重要的多自主体系统模型出发进行研究:生物系统中的群体模型、物理系统中的激光振列模型以及社会经济系统中的多人博弈模型,探讨当自主体的状态与网络的拓扑结构耦合在一起时多自主体系统集体行为分析的方法与规律。二是基于多自主体的分布式优化。我们将研究基于梯度的凸优化算法与一般指标函数的分布式优化算法设计与分析。三是多自主体系统的调控。我们将对几类不同的多自主体系统提出有效的干预策略并给出定量的理论研究。多自主体系统分析的方法和规律为分布式优化和调控奠定基础,合理有效的分布式优化算法与调控策略将为实际系统的优化与调控提供技术指导与理论支持。
中文关键词: 多自主体系统;集体行为;干预;分布式滤波;分布式控制
英文摘要: Complex System study is called "the 21st century's science", and Multi-Agent System is one of the attracting research fields in the Complex System study. Multi-agent systems arise from diverse fields, including biology, physics, engenieering and soci-economic systems. We will investigate multi-agent systems from three aspects: ① Analysis of the collective behavior of multi-agent systems. We start with our study from three basic and important multi-agent systems: flocking model in biological system, coupled laser arrays in physical systems and multi-player games in soci-economical systems。 We will explore the methods and laws of collective behavior of multi-agent systems in which the agents' states and the network topology are coupled together. ② Distributed optimization based on multi-agent systems. We will study the distributed gradient methods for convex optimization, and investigate the distributed optimization methods for general objective functions. ③ Regulation of multi-agent systems. We will propose and analyze the efficient interevention strategies for several classes of multi-agent systems. The methods and laws in the analysis of multi-agent systems will set up foundation for the optimization and regulation of multi-agent systems, and reasonable and efficient optimization and regulation methods will pr
英文关键词: multi-agent systems;collective behavior;intervention;distributed filtering;distributed control