基于DPSO最小碰集算法的掩盖故障识别

针对可测性分析中掩盖故障识别的难点,提出一种掩盖故障存在性的判定及计算其冲突集方法;并将离散粒子群算法(discrete binary particle swarm optimization,DPSO)用于求解冲突集的最小碰集,实现掩盖故障最小碰集的求解;为克服DPSO易陷入局部最优的缺点,还比较了惯性权重对DPSO算法性能的影响。实例验证表明:与求解掩盖故障的其他方法相比,惯性权重线性变化的DPSO算法不仅提高了算法效率,而且避免了其他算法求解时容易出现"计算爆炸"的问题,尤其适合于识别大型复杂系统的掩盖故障。

Minimal hitting sets algorithm of identifying masking false failure sets based on DPSO

Analyzing the masking false failure sets for a given fault is very difficult.A method of identifying the masking false failure sets and computing their conflicting sets is proposed;finding the minimal hitting sets algorithm based on DPSO is applied in enumerating the masking false failure sets;and compared the effect of DPSO with inertial weights because of its characteristic of getting into local optimization.The example in the paper shows: DPSO with linearly decreasing weight not only improves the efficiency but also avoids the phenomena "computation explosion" in contrasting to other similar algorithms,which is adapted to identify masking faults in large-scale compicated systems.