An Open Multi-agent Platform for Price Strategy Optimization of Generators In Market Environment

The generator’s competition strategy has become a pressing research field since the opening of power market. In real market situation Generators have to deal with capacity allocation among different markets (i.e. day-ahead spot market, contract market and ancillary service market) and competition strategy optimization simultaneously. However these two interrelated problems have been studied separated with little practical reference. The allocation of capacity among three markets sets the foundation for strategy optimization while actual market performance is the evaluation criterion of capacity allocation. The key of decision-making is the risk of price uncertainty and its manipulation. Our work provides a new realistic platform of competition strategy optimization for Generation Companies. The process of strategy optimization is completed in a multi-agent system that is interactive with the user. Different kinds of software agents are designed to fulfill the optimization function. Scenario tree generation algorithm is used to deal with the uncertainty of electricity price and genetic algorithm is used to solve the complex optimization problem of generation capacity allocation among different periods. Then competition strategies in spot as well as contract and ancillary market are defined. Finally evaluation process is designed for strategy improvement, which effectively combines the problem of generation capacity allocation with competition strategy optimization. Key-Words: Electricity market; Agent-based system; Scenario tree generation; Capacity allocation; Trading strategy; Power system.