基于自适应粒子群支持向量机的短期电力负荷预测

针对粒子群优化算法存在易陷入局部最优点的缺点,提出了一种新的基于平均粒距的自适应粒子群优化算法(ASPO).该算法利用种群多样性信息对惯性权重进行非线性调整,并在算法的后期引入速度变异算子和交换算子,使算法摆脱后期易于陷入局部最优点的束缚,同时又保持前期搜索速度快的特性.将该算法应用到基于支持向量机的短期电力负荷预测模型中,对支持向量机的参数进行优化.对某电网的短期负荷预测实际算例仿真分析表明,所提出的基于APSO-SVM方法的预测精度明显优于传统的SVM方法,且速度较快,因此,该算法用于短期电力负荷预测是有效可行的.

A Short-Term Load Forecasting Approach Based on Support Vector Machine with Adaptive Particle Swarm Optimization Algorithm

Aiming at the precocious convergence problem of particle swarm optimization algorithm,an adaptive particle swarm optimization algorithm(APSO) is presented,in which the inertia weight is nonlinearly adjusted by the information on population diversity and then the velocity mutation factor and position interchange factor are both introduced in to get rid of the constrict due to precocious convergence with fast hunting speed kept as previous.The algorithm has been applied to the optimization of parameters in SVM.A short-term load forecasting model based on SVM with adaptive particle swarm optimization algorithm(APSO-SVM) is thus presented.The simulation results showed that APSO-SVM can offer more accurate forecasting result than conventional SVM method.Therefore,the approach is efficient and practical to a short-term load forecasting of electric power system.