城市小时级需水量的改进型引力搜索算法--最小二乘支持向量机模型预测

Gravitational search algorithm–least squares support vector machine model forecasting on hourly urban water demand

中文关键词  智能控制  需水量预测  最小二乘支持向量机  改进的引力搜索算法

英文关键词  intelligent control  water demand forecasting  least squares support vector machine  ameliorated gravitational search algorithm

中文摘要

      本文研究利用最小二乘支持向量机(least squares support vector machine, LS--SVM)算法建立城市小时级需水量预测模型. 采取精英策略, 自适应的速度更新权重系数, 同时引入粒子历史最优信息对引力搜索算法(gravitational search algorithm, GSA)进行了改进. 最后采用改进型引力搜索算法(ameliorated gravitational search algorithm, AGSA)优化LS-SVM水量预测模型的正规化参数和核参数来提高模型的预测精度及预测速度. 理论测试与实例分析表明, 基于AGSA比基于GSA, 遗传算法(genetic algorithms, GA)和粒子群优化算法(particle swarm optimization, PSO)的LS--SVM水量预测模型具有更好的预测精度, 从而验证了基于AGSA的LS--SVM算法适用于小时级需水量预测问题, AGSA适用于多领域的模型参数的优化过程.

英文摘要

      We investigate the model of hourly urban water demand forecasting with least squares support vector machine (LS–SVM). The convergence performance of gravitational search algorithm (GSA) is improved by employing an elite strategy and an adaptive velocity with updated weighting factor. Furthermore, the historical optimal information is introduced to speed up the convergence of GSA. The ameliorated GSA, called AGSA, is adopted to optimize the regularization parameters and kernel parameters of LS–SVM used in the hourly urban water demand prediction model. Theoretical analysis and experimental results show that the AGSA-based hourly urban water demand forecasting model achieves better regression precision than other models respectively based on particle swarm optimization (PSO), genetic algorithms (GA), and GSA. This implies that AGSA-based LS–SVM algorithm can be successfully used to build the model of hourly urban water demand forecasting. In fact, AGSA can also be applied to process of parameter optimization in many other fields.