非Markov系统的可靠性仿真

Reliability Simulation of Non-Markovian System

在进行非Markov系统的可靠性仿真时,为获得准确的状态转移时间抽样,提出剩余分布的近似抽样(ASRD)方法,利用多个均匀分布拟合一般形式的剩余分布,通过拟合参数d调整抽样精度和抽样效率。在此基础上,提出基于强制转移(FT)策略的ASRD(FT-ASRD)仿真方法。多个算例结果表明,当5≤d≤10时,FT-ASRD相对于常规仿真方法具有更高的效率,而且仿真结果的相对误差较小。

In order to sample precisely the state transition time when simulating non-Markovian system,a new sampling method named Approximate Sampling for Residual Distribution（ASRD） method is put forward,in which several uniform distributions are used to approach general residual life distributions. The approach-coefficient d is used to adjust sampling precision and efficiency. Another method named Forced Transitions（FT） based ASRD（FT-ASRD） simulation method is brought forward to simulate high-reliable non-Markovian systems. FT-ASRD uses FT strategy to get the unreliability parameter. Several examples show that,when 5≤d≤10,FT-ASRD is more efficient and has lower relative error compared with crude simulation methods.