基于自适应遗传算法的多变量系统近似模型设计方法  

Approximate Model Design Method for Multivarible System Based on Adaptive Genetic Algorithm

作　　者：刘旭[1] 薛福珍[1] 唐雷[1] 

机构地区：[1]中国科学技术大学自动化系,合肥230027

出　　处：《化工自动化及仪表》2009年第1期27-30,共4页Control and Instruments in Chemical Industry

摘　　要：研究多变量系统近似模型设计方法中近似模型的选取问题,提出基于自适应遗传算法的最优近似模型选取算法。该方法是一种启发式智能选择方法,它改变了一直以来凭设计人员的经验选取近似模型的做法,使用方便,准确度高。所实现的基于近似模型技术的多变量系统鲁棒设计平台,集模型优化、系统分析和系统设计于一体,在工程上具有较大的应用价值。在自行研制的多变量系统设计平台中对一个参数不确定性工业对象的鲁棒系统进行设计,结果表明了该算法的有效性和实用性。The approximate model selection problem in multivariable system was studied. An optimal approximate model selection method based on self-adapting genetic algorithm was proposed. It was a heuristics intelligent selection method thus changed the way of selecting approximate model by the experience of designers. This method is convenient in using and has high accuracy. The multivariable system robust design platform based on approximate model technique has model optimization, system analysis and system design in one. In the self-developed muhivariable system design platform for a parameter uncertainty robust industrial object system design shows its effectiveness and practicality.

关 键 词：遗传算法 近似模型 多变量系统 鲁棒设计 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]