基于多智能体的用户分时电价响应模型

Modeling of User Response to Time-of-Use Price Based on Multi-Agent Technology

科学制定峰谷分时电价政策,有利于提高系统可靠性与经济性。基于多智能体技术,提出了智能体系统(multi-agent system,MAS)总体框图,阐述了供电智能体、用户智能体和政府部门智能体的学习过程及其相互影响,结合电量电价弹性矩阵,考虑了不同用户的用电和响应特点以及用户智能体内部的相互影响。仿真过程可以动态显示供电智能体的调整动作、用户智能体的响应行为和负荷越限情况,以及政府部门智能体的干预行为。仿真结果表明:根据文中模型获得的峰谷分时电价策略具有更好的"移峰填谷"效果,兼顾供需双方的经济利益和用户用电习惯,供电、用户、政府智能体间实现信息共享。

It is favorable to improve reliability and economy of power system to reasonable draft the policy of time-of-use （TOU）. Based on multi-agent technology, a block diagram of multi-agent system （MAS） is proposed, and the learning process of multi-agents for power supply branch, consumers and government department as well as the interaction among them are expounded. Combining with elastic matrices of electricity price and quantity of electricity, the features of power consumption and response of different consumers as well as the interaction within the multi-agent of consumers are taken into account. Simulation process can dynamically reveal the regulation within the multi-agent of power supply, the response behavior within the multi-agent of consumers and the out-of-limit load and the behavioral Intervention from the multi-agent of govemment department. Simulation results show that adopting the proposed TOU pricing strategy a better peak load shifting effect can be achieved; economic benefit of both sides and the power consumption habit of users can be simultaneously considered; the information sharing among power supply branch, consumers and government department can be implemented.