摘要
化工过程的动态优化,大多较为复杂,有相当的难度.新近发展的粒子群优化算法,基于群智能机理,适于求解连续问题,但它不具备遍历特性,影响了全局搜索能力.本文拟引入混沌机制,以混沌变量的遍历性改进粒子群算法,使其更全面地获取目标函数的有用信息,并反映到逐代更新的个体极值和群体极值中,可更有效地带领粒子群移向最优解,提高了全局搜优效率.由此构建为混沌粒子群算法,经多个性能测试,表明其搜索能力优于经典粒子群算法,引入混沌机制是有效的.将其用于Park-Ramirez生物反应器补料流率的动态优化,也取得了满意的效果.
Most of dynamic optimizations in chemical industry are complicated and difficult to solve. Particle swarm optimization algorithm （PSO） has been developed recently, which is suitable for solving continuous problem, but its lack of ergodicity impairs the algorithm global search property. In this paper, PSO was improved by introducing chaos to the algorithm. Taking advantage of chaos＇s ergodicity, PSO could comprehensively get the useful information about the objective function, which was reflected in the present global optimal point and the optimal point of each of particle in each iteration. By this, PSO eould more effectively adjust search direction of each particle and finally get the global optimal point of the problem. Experimental results showed that the proposed method was successful. The algorithm was applied to dynamie optimization of the feed-rate of a Park-Ramirez bioreactor and satisfying results were obtained.
出处
《化工学报》
EI
CAS
CSCD
北大核心
2006年第9期2123-2127,共5页
CIESC Journal
基金
国家自然科学基金项目(20276063).~~
