基于遗传算法的反例理解  

An algorithm for explanation counterexamples by use of a genetic algorithm

作　　者：李雅 黄少滨[1] 李艳梅[1] 迟荣华 郎大鹏 

机构地区：[1]哈尔滨工程大学计算机科学与技术学院,黑龙江哈尔滨150001

出　　处：《哈尔滨工程大学学报》2016年第10期1394-1399,1437,共7页Journal of Harbin Engineering University

基　　金：国家科技支撑计划课题(2012BAH08B00)

摘　　要：在复杂系统的错误定位过程中,为了寻找合适的证例,提出一种结合模型检测技术的基于遗传算法的反例理解算法。利用遗传算法,针对模型检测中的模型具有初始状态,状态之间有相应的迁移关系等特点,初始化种群、设计适应度函数并进行指向性的变异操作。实验表明该算法能快速有效的获得最近证例,帮助模型检测的反例理解。In order to seek the appropriate witness to help fault location of a complex system,a counterexample understanding algorithm in combination with a model checking technique on the basis of a genetic algorithm was proposed. The model used for model checking has an initial state,and the transfer relationship exists between the corresponding states. Therefore,a population was initialized. Also,a fitness function was designed and a directive mutation operation was conducted. Experimental results demonstrate that the algorithm can find the nearest witness rapidly and effectively and help the counterexample understanding of model checking.

关 键 词：模型检测 遗传算法 反例理解 错误定位 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]