基于BQPSO的多故障最小候选集生成技术研究  

Study on Generation of A Multiple Fault Minimal Candidates Set Based on BQPSO

作　　者：李宝晨[1] 黄考利 连光耀 吕晓明 张西山[1] 

机构地区：[1]军械工程学院,石家庄050003 [2]军械技术研究所,石家庄050003

出　　处：《计算机测量与控制》2013年第6期1476-1479,1482,共5页Computer Measurement &Control

摘　　要：多故障最小候选集生成是多故障诊断策略的首要问题,研究了利用二进制量子粒子群优化算法(Binary Quantum ParticleSwarm Optimization,BQPSO)生成多故障模糊组最小候选集的方法;首先利用紧急表示法描述某或节点上的多故障模糊组,其最小候选集即多故障模糊组的最小碰集;然后利用BQPSO算法求解多故障模糊组的最小碰集,通过进行二进制编码,并构造引导粒子搜索问题解的适应度函数,解决BQPSO算法求解最小碰集的适应性问题,并保证了算法尽可能搜索冲突集的全部碰集;最后通过某系统实例对算法的有效性进行了验证;结果表明,该算法具有全局搜索能力强、控制参数少、收敛性能好、运算简便等优势,能有效应用于多故障最小候选集问题的求解。The generation of a multiple fault minimal candidates set is the first step in making the multiple fault diagnostic strategy. The binary quantum particle swarm optimization algorithm is applied to generate the minimal candidates set. Firstly, the multiple fault ambiguity group in a certain OR node is described by using compact set notation, so the minimal candidate set is a minimal hitting set. Then, by using binary encoding and constructing fitness function which instructs swarm find problematic solutions, the BQPSO is applied in finding minimal hitting set successfully, and guarantees to find the hitting sets as many as possible. Finally, the experiment results of a certain real system verify the effectiveness of the algorithm. Experiment results indicate that this method has less control parameters, good convergence performance, and simple operation, can solve the minimal hitting set effectively.

关 键 词：多故障诊断 最小候选集 最小碰集 BQPSO算法 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置]