Electricity point price evaluation using hybrid algorithm

Estimation of price is the most crucial task and the basis for making decisions in competitive bidding strategies. Robustness, reliability and optimal profits for the market players are the main concerns which can be achieved by a point price forecasting module comprising of diminutive prediction errors, less computational time and reduced complexity. Hence in this work, an integrated approach based on Artificial Neural Networks (ANN) trained with Particle Swarm Optimization (PSO) is proposed for short term market clearing prices forecasting in pool based electricity markets. The proposed approach overcomes the difficulties like trapping towards local minima and moderate convergence rates as in existing methods. The work was deliberated on mainland Spain electricity markets and the results obtained are compared with hybrid models presented in the past literature. The response shows decrement in forecasting errors that are identified in price forecasting. The complete research may assist the ISO in finding out the key factors that are fit for prediction with low errors.