MMAS与粗糙集在齿轮箱故障诊断中的应用  

Application of MMAS and rough sets in fault diagnosis of gearbox

作　　者：沈仁发[1] 郑海起[1] 金海薇[2] 康海英[1] 张俊武 

机构地区：[1]军械工程学院火炮工程系 [2]装备指挥技术学院重点实验室 [3]65571部队

出　　处：《振动与冲击》2010年第11期190-194,共5页Journal of Vibration and Shock

基　　金：国家自然科学基金资助项目(50775219)

摘　　要：粗糙集理论是一种新的处理含糊和不确定性知识的数学工具,属性约简是粗集理论研究的重要内容,属性约简算法有很多种,而计算一个最佳约简是NP难问题。为了能够有效地获取信息系统的约简,提出了一种新的约简算法。该算法选择最大-最小蚂蚁系统(MMAS),以Fisher准则作为启发式信息来提高搜索效率,将蚁群优化算法引入属性约简中,利用粗糙集理论对故障诊断决策表进行约简,形成清晰、简明的故障诊断规则,为下一步的故障诊断打下了坚实的基础。Rough set theory is a relatively new soft computing tool to deal with vagueness and uncertainty.Attribute reduction algorithm is the key point of rough set study.There are many kinds of attribute reduction algorithm.However,it is proved that finding the best reduction is the NP-hard problem.For the purpose of getting the reduction of systems effectively,an improved algorithm was put forward.Max-min ant system (MMAS) was selected and the Fisher criterion was adopted as heuristic information to improve the searching efficiency.Simplifying the fault diagnosis decision table with the rough sets theory,then clear and concise decision rules could be obtained,they laid a foundation for the next fault diagnosis.

关 键 词：粗糙集理论 属性约简 最大-最小蚂蚁系统 决策规则 

分 类 号：TH165.3[机械工程—机械制造及自动化]