Study on Power System Load Forecasting Based on MPSO Artificial Neural Networks

Abstract

As a representative method of swarm intelligence particle swarm optimization (PSO) is an algorithm for search the multidimensional complex space through cooperation and competition among the individuals in a population of particles. A novel modified particle swarm optimization (MPSO) algorithm is proposed. The MPSO is determined by linearly decreasing inertia weight and constriction factor weight to speed up global search, also is combined with crossover and mutation to avoid the common defect of premature convergence. According to the different purpose of power system load forecasting, serial artificial neural network model to forecast power system short term load is introduced. Using the proposed MPSO algorithm we simulate the prediction of power system short load, the results shows that forecasting model based on MPSO artificial neural network algorithm can get a better forecasting effect compared with conventional BP algorithm. This approach reduces the training time and accelerates the speed of PSO algorithm and improves the adaptability of the artificial neural networks system