论文部分内容阅读
基于动态贝叶斯网络处理动态不确定性问题的过程中推理是非常重要的,而推理算法的优劣决定着推理的执行效率。该文提出一种较简单的112片联合树算法,在不需要限制消去顺序且只作一次扩展的条件下构造联合树,所以算法简单且具有较小的复杂度。
Inference in the process of dynamic uncertainty based on dynamic Bayesian network is very important, and the quality of inference algorithm determines the execution efficiency of reasoning. This paper proposes a simple 112-piece joint tree algorithm that constructs a joint tree without restricting the order of elimination and expanding it only once, so the algorithm is simple and has a low complexity.