Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Agent-based computational economics in power markets - multi-agent based simulation as a tool for decision support

: Veit, D.; Fichtner, W.; Ragwitz, M.

Andrysek, J.:
Multiple participant decision making
Magill: Advanced Knowledge International, 2004 (International series on advanced intelligence 9)
ISBN: 0-9751004-5-9
Book Article
Fraunhofer ISI ()
power market; renewable energy; agent-based computational economic (ACE); multi-agent system; simulation

The aim of this research in progress is to examine the effect of CO2-emission trading and the intensified application of renewable energy sources to the liberalized power market. The research approach applies methods from a combination of economics, natural sciences and computer science in order to obtain decision support for the planning of future power plant structures, investment decisions, and emission scales.
Besides the liberalization, future power markets are increasingly influenced by regulatory acts which are mainly responding to the necessity of a structured reduction of CO2-emissions. For CO2-eminssions, which are generated by the production of electric power, an emission trading system with obligatory participation is being launched from 2005 on. Since in the same period, the amount of power which has to be produced from renewable energy sources is increased dramatically, an additional incoming aspect is the fluctuating character of many of those sources.
These changed general conditions in the industrial power sector cause novel structures at international power markets and additionally a new emissions situation. The aim of this research is the conception of a methodologically innovative concept to simulate power markets based on multi-agent systems (MAS) and agent-based computational economics (ACE) which fulfils the described desiderata.
Since heterogeneous utility functions for individual market participants can be modelled - in comparison to existing centrally steered simulation methods - a multi-agent based simulation approach promises more realistic simulation results. Specific behaviour and the dynamic adoption of the strategies of individual participants based on the experiences derived from market behaviour can explicitly be simulated. The aim of the research approach is to obtain a strengthened understanding of the mutual effects among heterogeneous participants and the dynamics of the market of CO2-emissions as well as renewable energy resources. To disseminate these results, acting recommendations for participants at international power markets as well as political and regulatory authorities are derived.