A HYBRID PARTICLE SWARM OPTIMIZATION BACK PROPAGATION ALGORITHM FOR SHORT TERM LOAD FORECASTING

As accurate Short Term Load Forecasting (STLF) is very important for improvement of the management performance of the electric industry, various short term loads forecasting methods have been developed. This paper addresses an issue of the optimal design of a neural network based short term load forecaster. A new hybrid evolutionary algorithm combining the Particle Swarm Optimization (PSO) algorithm and Back Propagation (BP) algorithm, referred to as HPSOBP algorithm, is proposed to evolve the optimum large neural network structure, connecting weights and bias values for one-day ahead electric load forecasting problem. The hybrid algorithm can make use of not only strong global searching ability of the PSO algorithm, but also strong local searching ability of the BP algorithm. In addition, the input layer of the proposed ANN model receives all relevant information that contributes extensively to the prediction process. The proposed method is applied to STLF of the local utility. Data are clustered due to the differences in their characteristics. Special days are extracted from the normal training sets and handled, separately. In this way, a solution is provided for all load types, including working days, weekends and special days. The experimental results confirm that the proposed method optimized by HPSOBP can quicken the learning speed of the network and improve the forecasting precision compared to the BP and PSO methods and gives lower percent errors all the time. Thus, the proposed method is practical and effective for STLF problem and can be applied to automatically design an optimal load forecaster based on historical data.