Optimizing of SVM with Hybrid PSO and Genetic Algorithm in Power Load Forecasting

In this paper, we propose Hybrid Particle Swarm Optimization (HPSO) with genetic algorithm(GA) mutation to optimize the SVM forecasting model. In the process of doing so, we first use HPSO with genetic algorithm to make pretreatment of the data. PSO with GA model is a method for finding a solution of stochastic global optimizer based on swarm intelligence. Using the interaction of particles, PSOGA model searches the solution space intelligently, this will find out the best one and reduce the redundant information. The PSOGA-SVM method proposed in this paper is based on the global optimization of PSOGA and local accurate searching of SVM. And to prove the effectiveness of the model, single SVM algorithm and BP network was used to compare with the result of PSOGASVM. The results show that the model is effective and highly accurate in the forecasting of short-term power load than the other models. The root-mean-square relative error (RMSRE) of new model is only 1.82%, the SVM model and BP network is 2.43%, 4.10% separately.