This paper proposes an agent-based computational approach to study physical constrained electricity markets. The computational model consists of repeated day-ahead market sessions and a two-zone transmission network. Different inelastic load serving entities configurations are considered for studying how producers learn to strategically decommit their units and how they exercise market power by profiting from transmission network constraints. Learning producers are modeled by different multi-agent learning algorithms, such as the Q-Learning, the EWA learning and the GIGA-WoLF. Computational results point out that all learning models considered are able to learn to appropriately decommit their units and to sustain the exertion of zonal market power. ©2008 IEEE.
Supply-side gaming on electricity markets with physical constrained transmission network
GUERCI, ERIC;CINCOTTI, SILVANO;DELFINO, FEDERICO;PROCOPIO, RENATO;
2008-01-01
Abstract
This paper proposes an agent-based computational approach to study physical constrained electricity markets. The computational model consists of repeated day-ahead market sessions and a two-zone transmission network. Different inelastic load serving entities configurations are considered for studying how producers learn to strategically decommit their units and how they exercise market power by profiting from transmission network constraints. Learning producers are modeled by different multi-agent learning algorithms, such as the Q-Learning, the EWA learning and the GIGA-WoLF. Computational results point out that all learning models considered are able to learn to appropriately decommit their units and to sustain the exertion of zonal market power. ©2008 IEEE.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.