融合分层抽样和动态抽样的多状态网络可靠度M-C估计算法

M-C Estimation Algorithm for Multistate Network Reliability Based on Fusion of Hierarchical Sampling and Dynamic Sampling

基于多状态网络可靠度的Monte-Carlo(M-C)估计算法,考虑融合分层抽样和动态抽样的M-C估计算法.先在基于状态树搜索分层抽样方法的基础上,通过设定概率阈值α改变分层原则,使分层抽样便于实现;再利用动态抽样,在产生网络无效状态时动态生成网络各边的容量值,从而不需对所有边进行抽样即可产生无效网络状态,缩短了仿真时间.仿真实验表明,动态抽样能缩短仿真时间,但优势会随着网络可靠度的增大而逐步消失,较适用于可靠度低的多状态网络.

Based on a Monte-Carlo（M-C）estimation algorithm of multi state network reliability,the author considered an M-C estimation algorithm based on fusion of hierarchical sampling and dynamic sampling.The hierarchical sampling was realized by setting the probability thresholdαto change hierarchical principle based on the hierarchical sampling method of state tree search.Using dynamic sampling,the capacity value of each side of the network was dynamically generated when the network was invalid,so that the invalid network state could be generated without the sampling of all edges,and the simulation time was shortened.Simulation results show that the dynamic sampling can shorten the simulation time,but the advantage will gradually disappear with the increase of the network reliability,and it is more suitable for the multi state network with low reliability.