A multi-agent trading platform for electricity contract market

An agent-based negotiation platform for power generating and power consuming (purchasing) companies in contract electricity market is presented. An intelligent agent implements the negotiation process by selecting a strategy based on learning algorithm in an interactive manner with the user. Two kinds of learning algorithm - fuzzy logic controller modification of basic genetic algorithm for negotiation strategy optimization, and reinforced learning algorithm for negotiation tactics parameter modification - are provided for the agent. Protocol operation semantics that is flexible and can handle sequential message exchange is used as agent communication mechanism. The paper presents the architecture and negotiation strategy of agents. The software implementation of the platform is discussed in detail.