基于改进全局和声搜索算法LSSVM的短期电力负荷预测

摘要

针对全局和声搜索算法(GHS)存在的缺陷,提出改进全局和声搜索算法(IGHS).该算法通过扩大最优和声搜索区域,并在搜索过程中引入受和声库影响的微调变量,从而增强了算法跳出局部极小值束缚的能力.将该改进算法应用于电力负荷预测中,提出基于改进全局和声搜索算法最小二乘支持向量机(IGHS-LSSVM)的负荷预测方法.利用某电力公司的历史数据进行仿真,结果表明方法具有较高的预测精度.%To overcome the drawback of the global harmony search(GHS) algorithm, the improved global harmony search algorithm(IGHS) is proposed. By means of broadening the searching area in the GHS algorithm and introducing fine ?tuning variable affected by the harmony memory, the search efficiency is improved. Then, the proposed IGHS algorithm is applied to the optimization of parameters in LSSVM short-term load forecasting. The load data and meteorological data are utilized to test the forecasting model, and the simulation results show that the proposed method can improve the predicting accuracy.