The ant colony optimization (ACO),the genetic algorithm (GA),and the particle swarm optimization (PSO) have different strengths and weaknesses.In order to fully use the fine global search capabilities of both ACO and GA and the hierarchical search mechanism of PSO,both the hybridizing ant colony and particle swarm optimization (named ACO + PSO) and hybridizing genetic algorithm and particle swarm optimization (called GA + PSO) based least square support vector machines (LSSVM),referred respectively to as ACO + PSO-LSSVM and GA + PSO-LSSVM,have been proposed to predict the non-Gaussian fluctuating wind velocity.Subsequently,the non-Gaussian wind velocity of a super-tall building was forecasted by using the ACO + PSO-LSSVM and the GA + PSO-LSSVM.For the purpose of comparison,the results for the non-Gaussian wind velocity of ACO-LSSVM,PSO-LSSVM,and GA-LSSVM were provided simultaneously.Through scrutinizing the predicted non-Gaussian wind velocity time-history,predicted values of correlation functions,and predictive performance evaluation indices,it has been shown that the two proposed hybridizing intelligent optimization based LSSVM algorithms possess better accuracy and higher robustness for the prediction of nonGaussian fluctuating wind velocity of super-tall buildings.%考虑智能优化:蚁群算法(ACO)、遗传算法(GA)和粒子群算法(PSO)各自优缺点,并为充分发挥蚁群、遗传算法较好的全局搜索能力和粒子群算法的分级搜索机制,提出混合蚁群和粒子群优化(ACO+ PSO)和混合遗传算法和粒子群优化(GA+ PSO)最小二乘支持向量机(LSSVM)的非高斯脉动风速预测模型,分别称为ACO+ PSO-LSSVM和GA+ PSO-LSSVM.运用ACO+ PSO-LSSVM和GA+ PSO-LSSVM预测模型对某超高层建筑的非高斯脉动风速进行了预测;为比较目的,同时给出ACO-LSSVM、PSO-LSSVM和GA-LSSVM的非高斯脉动风速预测结果.经仔细检查非高斯脉动风速时程预测值、相关函数预测值以及预测性能评价指标,验证了基于混合智能优化LSSVM对非高斯脉动风速预测的有效性和优势.
展开▼
