Coevolutionary Genetic Fuzzy System to Assess Multiagent Bidding Strategies in Electricity Markets

In this paper we study a genetic fuzzy system approach to assess suitable bidding strategies for agents in online auction environments. Assessing efficient bidding strategies allows evaluation of auction models and verification whether the mechanism design achieves its goals. Day-ahead electricity auctions are particularly explored to give an experimental instance of the approach developed in this paper. Previous works have reported successful fuzzy bidding strategies developed by genetic fuzzy systems and coevolutionary algorithms. Here we review the coevolutionary algorithm and present recent results of the bidding strategies behavior. We analyze how the evolutionary strategies perform against each other in dynamic environments. Coevolutionary approaches in which coevolutionary agents interact through their fuzzy bidding strategies permit realistic and transparent representations of the behavior of the agents in auction-based markets. They also improve market representation and evaluation mechanisms. Experimental results show that coevolutionary agents can enhance their profits at the cost of increasing system hourly price paid by demand, an undesirable outcome from the perspective of the buyers.