多智能体协调控制的演化博弈方法
EVOLUTIONARY GAME THEORETIC APPROACH TO COORDINATED CONTROL OF MULTI-AGENT SYSTEMS
提出一种基于演化博弈理论的多智能体系统协调控制方法. 在所建立的数学模型框架中智能体根据其自身的利益, 通过局部交互, 在博弈竞争中学习, 自主调整其行为. 根据系统整体性能的要求, 通过选择合适的博弈类型、设计适宜的收益计算方法、更新进化规则等, 实现对多智能体系统的控制. 在演化过程中, 无需指定某些特定个体的具体动力学行为, 只需通过种群的自适应进化即可实现整体目标. 以分工合作问题为例, 详细解析所提出的控制方法, 通过理论分析和仿真验证该方法可以实现多智能体系统的自适应协调控制.
In this paper, we propose an evolutionary game theoretic approach to coordinated control of multi-agent systems. In such fundamental mathematical framework, agents play games with their neighbors, and update strategies through local interaction. The agents autonomously regulate their behavior, according to their own benefit. % which is most like unsupervised study for individuals. Based on certain performance of the whole system, the control target is realized through selecting proper game type, designing fitness calculation and evaluation method, interactive restriction, update rule, and so on. In the evolutionary process, we do not need to design certain agents' dynamics and assign them what to do, but realize the final target through the self-adaptive evolution of the system. Taking division and cooperation problem as an example, we explain the proposed control approach in detail. The adaptive coordinated control process, which is a controllable, intelligent, and autonomous decision process, is analyzed and verified through simulation.
演化博弈理论 / 多智能体系统 / 协调控制. {{custom_keyword}} /
/
〈 | 〉 |