Forecasting the long-term electricity demand in Taiwan with a hybrid FLR and BPN approach

Precisely and accurately predict the electricity demand is an important task for the government in each country. In addition, establishing the lowest upper bound for the electricity demand also avoids unnecessary power plant investment. To this end, a hybrid fuzzy linear regression (FLR) and back propagation network (BPN) approach is proposed in this study. In the proposed methodology, multiple experts construct their own FLR equations to predict the future electricity demand from various viewpoints. Each FLR equation can be fitted by solving two equivalent nonlinear programming problems, based on the opinions of experts. In order to aggregate these fuzzy electricity demand forecasts, a two-step aggregation mechanism is used. First, fuzzy intersection is applied to aggregate the fuzzy electricity demand forecasts into a polygon-shaped fuzzy number, in order to improve the precision. Subsequently, a BPN is constructed to defuzzify the polygon-shaped fuzzy number and generate a representative/crisp value, in order to enhance the accuracy. The actual case of Taiwan is used to evaluate the effectiveness of the proposed methodology. According to the experimental results, the proposed methodology improved the precision and accuracy of the electricity demand forecasting by 33% and 99%, respectively.