基于改进CHC算法的模糊诊断规则优化方法研究  

A Study of the Optimization for Fuzzy Diagnostic Rules Based on the Reformative CHC Algorithm

作　　者：齐怡[1] 秦红磊[1] 沈士团[1] 李驿华[1] 

机构地区：[1]北京航空航天大学电子信息工程学院,北京100083

出　　处：《航空学报》2004年第4期362-367,共6页Acta Aeronautica et Astronautica Sinica

摘　　要：在故障诊断专家系统的自学习中,同时需要学习大量的模糊规则,造成规则优化算法的运行时间过长,难以满足实际的需要。为了解决这个问题,将特征片断优势重组算子引入CHC算法中,并将之应用到规则优化领域。仿真证明,改进的CHC算法在模糊诊断规则优化方面能以比较短的计算时间达到预期的优化效果,明显优于国内其它改进的遗传算法和CHC算法,具有比较大的实用价值。In the machine learning of rules for the diagnostic expert systems, a mass of fuzzy rules have to be optimized one time. It makes the optimization algorithm run much time that can't accord with the need in practice. In order to solve this problem, a new genetic arithmetic operator-best fraction recombination operator is proposed, and the CHC algorithm is reformed with it, which is used in the optimization algorithm in this paper. The simulation shows that the reformative CHC algorithm has much advantage in the optimization for fuzzy rules because of getting the satisfying result in a short consumed time. It can solve the optimization of fuzzy diagnostic rules while the ordinary genetic algorithm cannot.

关 键 词：模糊规则 自学习 遗传算法 优势重组算子 

分 类 号：V267[航空宇航科学与技术—航空宇航制造工程] TP273.4[自动化与计算机技术—检测技术与自动化装置]