化工学报2009，Vol.60Issue(12)：3052-3057,6.
基于差分进化粒子群混合优化算法的软测量建模
Soft sensor modeling based on differential evolution-particle swarm optimization based hybrid optimization algorithm
陈如清1
作者信息
- 1. 嘉兴学院机电工程学院,浙江,嘉兴,314001
- 折叠
摘要
Abstract
In the process of ethylene production, ethylene yield cannot be measured on-line via traditional approaches. To resolve this problem, a novel differential evolution (DE) -particle swarm optimization (PSO) based hybrid optimization algorithm (DEPSO) was proposed. Then a soft sensor model for real-time measuring ethylene yield was constructed. The procedure of optimization was divided into two phases and the particles were divided into two sub-swarms, one sub-swarm searched via PSO and the other searched via DE at the same time. Evolution speed factor was introduced in judging local convergence of algorithm during the process of iteration, with two sub-swarms exchanging information in each iteration to avoid local optimum. Optimization test on several complex functions with high-dimension indicated that the improved algorithm performed better than standard PSO and DE in whole optimization capability. Application results showed that the soft sensor model based on the improved algorithm had high measurement precision as well as good generalization ability.关键词
乙烯收率/软测量建模/差分进化算法/混合优化算法Key words
ethylene yield/ soft sensor model/ differential evolution/ hybrid optimization algorithm分类
信息技术与安全科学引用本文复制引用
陈如清..基于差分进化粒子群混合优化算法的软测量建模[J].化工学报,2009,60(12):3052-3057,6.
