基于多agent的电力系统主从递阶恢复决策

Multi-agent Based Stackelberg Decision for Power System Restoration

将多agent技术的分布式信息处理和主从递阶决策方法有机结合,提出一种分布式电力系统恢复的主从递阶决策模型。采用面向agent技术设计了主从递阶恢复决策系统的体系结构、各级调度agent的内部功能结构以及基于承诺和约定的交互协作机制。各级调度层的承诺为其恢复行为提供了一定程度的可预见性,约定作为变化的恢复过程中监控子调度层承诺进展的手段,为动态环境中的主从决策层的协作提供了灵活性。仿真结果显示,该模型不仅能够处理分布式群体决策中的多决策主体之间的配合问题,而且能灵活处理恢复过程中所出现的问题。

By integrating multi-agent technology with Stackelberg decision theory, a leader-follower hierarchical decision model is proposed for distributed restoration of a power system after blackout. The distributed decision-making system architecture is constructed with agent-oriented technology. Its internal function module is designed for all levels of the dispatching agent. The coordination mechanism of different agents is based on commitments and conventions. From the commitment, the behavior of dispatching is foreseen. The convention, which is susceptible to the accident during the restoration process, provides the flexibility for cooperation between the leader and followers in the dynamic environment. The simulation results show that the model proposed can not only effectively cope with the coordinate dispatching in the distributed decision-making process, but also flexibly deal with the problems encountered in the restoration process.