Application of an Adaptive Network-Based Fuzzy Inference System Using Genetic Algorithm for Short Term Load Forecasting

This paper discusses a method to forecast short term electricity load using genetic algorithm (GA) optimized Adaptive Network-based Fuzzy Inference System (ANFIS). The structure and parameters of the adaptive fuzzy neural network are synchronously optimized using an improved genetic algorithm. A fitness function is applied to guide the search process which makes the searching more efficient. The speed of convergence is significantly accelerated without causing any instability. After well trained, the fuzzy neural network is used to analyze relevant factors influencing load prediction. The results show that the proposed genetic algorithm optimization of adaptive fuzzy neural network has a higher forecasting accuracy and requires a shorter training time than the artificial neural network (ANN) which makes it attractive and promising in practical applications.