Abstract: Many systems with high reliability requirements are usually designed with k-out-of-n:G structure.However,the reliability optimization for such systems is a rather difficult problem.To further expand the application,a new reliability optimization model using mixed redundancy strategy was proposed.The model introduced mixed redundancy strategy in the reliability optimization of k-out-of-n:G system for the first time,where any one of the (active,cold-standby,or mixed) redundancy strategy might be selected for each subsystem.The mathematical model based on continuous time Markov chain was developed for the accurate calculation of system reliability for the k-out-of-n:G system.A redundancy allocation problem and an engineering case problem were proposed,and a pseudo-parallel genetic algorithm was designed for solving.To evaluate the performance of the new model,the numerical analysis was performed.The experiments results showed that the proposed new model could obtain higher system reliability value compared with previous models and had a good application prospect.

Key words: k-out-of-n:G system, mixed redundancy strategy, redundancy allocation problem, continuous time Markov chain, pseudo parallel genetic algorithm

摘要： 很多对可靠性要求极高的系统通常被设计成k-out-of-n:G结构,然而对该类系统的可靠性优化是一个相当困难的问题。为进一步扩展模型的适用性,提出一种采用混合冗余策略的可靠性优化新模型。该模型首次在k-out-of-n:G系统的可靠性优化中引入混合冗余策略,可以为每个子系统选择(积极、冷备份或混合)冗余策略中的任何一个。首先,基于连续时间马尔可夫链为k-out-of-n:G系统建立精确计算系统可靠性的数学模型。其次,提出一个冗余分配问题和一个工程案例问题,并设计一种伪并行遗传算法进行求解。最后,为评估新模型的性能,通过对提出的两个问题进行数值分析来评估新模型的性能。实验证明,相比以往模型研究,所提新模型得到更高的系统可靠性值,具有良好的应用前景。

关键词: k-out-of-n:G系统, 混合冗余策略, 冗余分配问题, 连续时间马尔可夫链, 伪并行遗传算法