基于抽样的因果图仿真推理算法研究

摘要

由于因果图的经典推理的计算复杂度是NP难的,因此其不便于推广和使用.基于因果图理论和Monte Carlo思想,提出了基于抽样的A-R Sampling和重要性抽样的因果图仿真推理算法.在故障诊断中的有效应用验证了方法是可行的.%Because the causality diagram's reasoning calculation is a NP problem,it is inconvenient in its popularization and application. This paper presented two simulation reasoning algorithms including A-R sampling and importance sampling based on causality diagram. The effective application in fault diagnosis shows the method is feasible.