Agent-based simulator for the German electricity wholesale market including wind power generation and widescale PHEV adoption

Abstract

An agent-based model is applied to model the German electricity wholesale market with its four major German utility companies. The model is utilized to assess base and peak power spot prices for scenarios implying doubling or tripling wind generation capacity in Germany. Furthermore, the effect of 8 million Plug-In Hybrid Electric Vehicles (PHEVs), incorporating different charging/discharging patterns, on spot prices is evaluated. In the model the power generating units within the utilities are modeled by agents. These agents are trained to increase their profits by using a reinforcement learning approach combined with a genetic algorithm resulting in heuristically optimized bidding strategies. This approach allows to take into account strategic market behavior and the exercise of market power when analyzing future wind expansion and wide scale PHEV adoption scenarios. The wind generation is considered as an exogenous input to the model which estimates potential electricity prices and total cost for consumers.