Provision of an Optimal Strategy to Forecast the Prices Set by the Electricity Market in the Competitive Iranian Energy Market in Fall

Given the complexities of the electricity market, various factors, such as uncertainties, the ways upon which the markets are set, how the debts are settled, the market structure and regulations, production prices, constraints governing the units and networks, etc. are influential in determining the optimal pricing strategies. Various methods and models have been presented to resolve the pricing issue in the competitive electricity industry. The most prominent of which include pricing methods are based on the prediction of competitors’ behavior; also pricing methods based on the forecasts of market price, methods based on the game theory and lastly, pricing methods based on the intelligent algorithms. Therefore, this study was conducted to provide an optimal strategy in order to forecast the electricity market price set in the competitive Iranian electricity market (based on the data collected). In this paper, the proposed method uses a compound network based on the neural networks. The analyzed data include the amount of the consumed energy as well as temperature (if applicable) and the price set for the past days and weeks. The self-organizing map (SOM) network was used for the input clustering based on the similar days. A number of multilayer perceptron (MLP) neural networks were used to combine the extracted data consisting of the energy levels, the price set, and temperature (if possible). The results showed improvements in the performance of the smart systems based on the neural networks in predicting the electricity prices.