LAUSR

In the context of agent-based simulation framework of collusion, this paper seeks for two-sided tacit collusion among supply-side and demand-side participants in a constrained network and impacts of this collusion on the market outcomes. Tacit collusion frequently occurs in electricity markets due to strategic behavior of market participants arose from daily repetition of energy auctions. To attain detailed analysis of tacit collusion, state-action-reward-state-action (SARSA) learning algorithm and the standard Boltzmann exploration strategy based on the Q-value are used to model market participants’ behavior. A model is presented that integrates exploration and exploitation into a single framework, with the purpose of tuning exploration in the algorithm. In order to appraise the feasibility of collusion, a theoretical study on a three-node power system with three scenarios is depicted considering three Gencos and two Discos which proves the formation of two-sided tacit collusion between Genco and Disco. Simulation results show different collusive strategies of participants and how parameters of the algorithm impact on simulation outcomes. It is also shown that congestion on transmission line has a significant influence on behavior of market participants.