因果图迭代推理算法研究

针对因果图精确推理是NP(nondeterministic polynomial)难的问题，提出了一种迭代推理方法。首先,从图论的角度分析了因果图推理中概率计算的机理，并提出了矩阵解环的方法。在此基础上提出了一种迭代推理算法，该算法只需要进行简单的矩阵运算，大大简化了传统因果图推理复杂的计算过程，可以在多项式时间复杂度内实现推理。其次，分析了算法存在的问题并提出了改进的方向。最后，运用实例分析验证了该算法实现因果图推理的效果。研究表明，该算法能够有效地进行因果图推理，推理效率高，推理结果正确,为因果图提供了一种高效的近似推理方法，对因果图的应用具有重要意义。

Iterative reasoning algorithm of causality diagram

An iterative reasoning method is proposed to solve the problem that accurate reasoning of causality diagram (CD) is nondeterministic polynomial (NP) hard. By the point of view of graph theory , the probability computing mechanism of causality diagram reasoning is firstly analyzed, and a method of breaking down circuits by matrix is introduced. Then, an iterative reasoning algorithm which only needs simple matrix operations is proposed to greatly simplify the complex computing process of conventional causality diagram reasoning, thus achieving reasoning in the polynomial computation time complexity. Secondly, the problems and improved directions of the iterative algorithm are analyzed. Finally, an example demonstrates the effect of the iterative reasoning algorithm of CD. The research shows that the iterative reasoning algorithm of CD is effective and fast to work out right results.