Adaptive Learning For Strategic Bidding In Electricity Spot Market Based On Price Prediction

How to ensure the effectiveness of power producers’ bidding strategies is worth studying in the area of the electricity market.In an electricity market,power producers operate differently.They can bid freely in the market,and their profits are determined by their own bidding choices rather than the regulatory authority.In order to ensure that power producers can obtain sufficient generation revenue and recover investment costs,so as to maintain a reliable capacity of the system,it is necessary to study how power generation companies can ensure sufficient revenue through bidding strategies.Bidding strategy is a tool to help power producers participate in market bidding.Based on market information,power producers can use bidding strategies to estimate the market environment and make bidding decisions.However,in a uniform pricing electricity market,the supply and demand are changeable,and the clearing results are difficult to be predicted accurately.If the bidding strategy is unadvanced,there may be problems such as low profits.This study focuses on the bidding strategy for the traditional electricity market.First of all,an adaptive learning scheme for strategic bidding is proposed to ensure considerable effectiveness.Then,for extended research,the effects of the bidding strategy in different environments are taken into account.The main works are as follows:(1)A simulation model of the electricity market based on multi-agent technology has been established.The model includes multiple modules such as the producer models and the clearing model.It truly simulates the bidding process of producers in power spot trading,and thus provides a basis for producers’ bidding strategies.(2)An adaptive learning scheme for strategic bidding is proposed.The proposed scheme is based on an ensemble technique,where several machine learning algorithms based on different theories serve as the underlying algorithms,ensuring that certain algorithms are always effective in any scenario,and they will lead the bidding process.The key point of the bidding strategy is electricity price forecasting,and the algorithm will use electricity price forecasting to estimate the market environment,and then give recommended biddings based on the evaluation results.Taking into account the characteristics of the electricity market,this paper improves the price prediction model to obtain higher prediction accuracy.(3)The influence of market design on the profits of power producers is analyzed to study if they can limit market power.First,in terms of pricing mechanism,uniform pricing and pay-as-bid pricing are compared.Second,long-term transactions are introduced to mitigate the impact of producers’ market power.Third,the bidding strategies of power producers are upgraded and the market clearing result under this circumstance is studied.Simulation studies are presented to demonstrate the effectiveness of the proposed scheme,and the results show that adopting the adaptive learning bidding strategy may lead to more reasonable bidding behaviors and higher profits.The research of this paper can be further developed into a simulation platform to provide auxiliary decision-making for regulatory agencies,system operation agencies or market entities,which is beneficial to explain the market bidding behavior of power producers and help them make more reasonable predictions of market trends.