Reinforcement Learning Based on the Bayesian Theorem for Electricity Markets Decision Support

摘要

This paper presents the applicability of a reinforcement learningalgorithm based on the application of the Bayesian theorem of probability. Theproposed reinforcement learning algorithm is an advantageous andindispensable tool for ALBidS (Adaptive Learning strategic Bidding System), amulti-agent system that has the purpose of providing decision support toelectricity market negotiating players. ALBidS uses a set of different strategiesfor providing decision support to market players. These strategies are usedaccordingly to their probability of success for each different context. Theapproach proposed in this paper uses a Bayesian network for deciding the mostprobably successful action at each time, depending on past events. Theperformance of the proposed methodology is tested using electricity marketsimulations in MASCEM (Multi-Agent Simulator of Competitive ElectricityMarkets). MASCEM provides the means for simulating a real electricity marketenvironment, based on real data from real electricity market operators.