Back-testing the agent-based model AMIRIS for the Austrian day-ahead electricity market

Abstract

The energy transition requires significant changes to current energy systems. Especially electricity markets are in the spotlight of policy-makers, investors, and researchers. This is due to the emergence of new market participants and innovative technologies disrupting the dominance of conventional power supply. Thus, we present the open agent-based electricity market model AMIRIS, which allows investigating current and future electricity systems. In this work, we apply AMIRIS to simulate the Austrian day-ahead electricity market prices. We use but freely available data in hourly resolution for the year 2019 and perform a back-test of the model with historical prices. The results show a high level of agreement of simulated results and historical data with regard to statistical characteristics (e.g. average price and price duration curve). However, AMIRIS tends to overestimate lower prices and underestimate higher prices. Also, AMIRIS does currently not include strategical bidding components and looking at the price time series, differences between the simulated and historical values are apparent. We conclude that the flexibility and the convenient parameterization of AMIRIS make it a powerful tool to assess today's and tomorrow's research questions in the field of energy economics. However, for deeper insights on the electricity market, further research is required to integrate bidding strategies of, e.g. energy storage system operators.