基于Rollout算法的多值属性系统诊断策略

Diagnosis strategy for multi-value attribute system based on Rollout algorithm

针对多值属性系统的诊断策略问题,根据Rollout算法可改善基础启发式算法的特点,用Rollout算法对基于信息熵的多值属性系统诊断策略进行优化,提出一种新的多值属性系统诊断策略.理论和实验分析表明,Rollout算法的计算结果优于信息熵算法,计算时间在可接受范围之内,可用于复杂多值属性系统的最优诊断策略设计.

The problem of diagnosis strategy for multi-value attribute system is considered. Based on the property that Rollout algorithm can improve the base heuristic algorithm, Rollout algorithm is used to improve the information entropy based diagnosis strategy for multi-value attribute system. Then, a new diagnosis strategy for multi-value attribute system is proposed. The theory and experiments demonstrate that, the method is much better than information entropy algorithm and its computing time is acceptable, which can be used to design the optimal diagnosis strategy for complicated multi-value systems.