摘要
利用混沌映射的随机性和遍历性,将其引入粒子群优化算法,以提高算法的全局寻优能力,同时引入优进策略,以改善其局部寻优效率,在此基础上构建了混沌混合粒子群优化算法(CHPSO)。高维复杂函数的仿真优化试验表明,CHPSO全局寻优能力强、优化效率高。针对常规算法训练神经网络容易早熟收敛和陷入局部极值点的不足,采用CHPSO训练人工神经网络,由此构建CHPSO-NN模型,并应用于乙酸己酯催化酯化反应条件的建模,与BP-NN相比,其预测能力和稳健性都有较大提高,效果良好,与传统方法相比有明显的优越性。
a new algorithm, which is named as chaotic hybrid particle swarm optimization algorithm （CHPSO） , is proposed. CHPSO integrates chaotic mechanism for its ergodicity, stochastic property, and regularity, which enhance the global exploitation. And also integrates the eugenic strategy to improve the local exploration. Simulation results on high dimensional complex functions shows that CHPSO has powerful global optimization ability and high optimization efficiency. Then CHPSO was proposed to training Neural Networks, named CHPSO-ANN, to establish reaction model for esterification of hexyl acetate, the predict ability and stability of results have a quite increase compared with the traditional BP-ANN method.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2007年第8期1073-1077,共5页
Computers and Applied Chemistry
基金
浙江省自然科学基金项目(Y404082)
浙江省教育厅重点科研计划项目(20030836)
关键词
混沌
优进
粒子群优化
神经网络
酯化
chaotic, eugenic, particle swarm optimization, neural networks, esterification
