Optimal bidding strategy for GENCOs using reinforcement learning process based on the PAB model

Abstract

The electricity market with open access enables participants to gain more profit out of the bidding strategy. Every supplier tries to maximize its profit as a player in the market. The decision-making process of suppliers and their mutual performance in the market is a complicated problem, can be studied by modeling single-generator and multi-generator companies. The present paper proposes a model based on the reinforcement learning algorithm, is capable of making decisions for suppliers in the single — generator and multi-generator states on proposing a bidding strategy and simulating market outputs based on mutual actions. Hence, a comparison has carried out to examine the performances of generators in the single-generator and multi-generator states without considering constraints and by considering the effect of network constraints, which can impose considerable limitations on electricity markets. The market clearing mechanism is based on Pay As Bid (PAB) model, can be used to define the optimal bidding strategy for each supplier, find market balance and assess market performance. The proposed model has applied to the Nord Pool market and its effect has indicated.