arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1207.0097eess.SYcs.SY
收藏

允许基于选择的行为的多智能体系统中的合作目标实现

Cooperative Target Realization in Multi-Agent Systems Allowing Choice-Based Actions

the Chinese University of Hong Kong(香港中文大学);Dalian University of Technology(大连理工大学)

Ge Guo, Wing Shing Wong, Zhongchang Liu

AI总结:

本文针对基于初始选择的目标函数和控制的线性多智能体系统,利用变分法和拉格朗日乘子法,提出了最小化平均控制成本的最优控制设计方法,并扩展至任意目标矩阵。

摘要:

在本文中,我们研究了合作多智能体系统,其中目标目标以及智能体实施的控制依赖于它们在初始系统时间做出的选择。此类系统已在最近发表的几篇论文中进行了研究,主要从系统分析的角度探讨了控制通信复杂度、控制能量成本以及目标函数实现的可行性等问题。本文通过为基于分布式选择协同操控线性系统的多个智能体开发最优控制设计方法,延续了这一研究方向。对于满足特定结构约束的目标矩阵,我们推导出能够以最小控制成本实现指定目标的控制算法。我们比较了状态反馈和开环控制策略在目标实现中的表现,并将最优性结果扩展到任意目标矩阵。通过现代变分理论和拉格朗日乘子法,在指定目标状态约束集下最小化平均控制成本,得到了最优控制解。

英文摘要:

In this paper, we study cooperative multi-agent systems in which the target objective and the controls exercised by the agents are dependent on the choices they made at initial system time. Such systems have been investigated in several recently published papers, mainly from the perspective of system analysis on issues such as control communication complexity, control energy cost and the feasibility of realization of target functions. This paper continues this line of research by developing optimal control design methodology for linear systems that are collaboratively manipulated by multiple agents based on their distributed choices. For target matrices that satisfy particular structural constraints, we derive control algorithms that can achieve the specified targets with minimum control cost. We compare state-feedback as well as open-loop control strategies for target realization and extend the optimality result to an arbitrary target matrix. The optimal control solutions are obtained by minimizing the average control cost subject to the set of specified target-state constraints by means of modern variation theory and the Lagrange multiplier method.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑