Component reassignment for reliability optimization of reconfigurable systems considering component degradation

This paper aims to optimize the lower bound of the reconfigurable system reliability by reassigning the degrading components, which is common in practical engineering systems. First, considering the reconfigurable linear consecutive k-out-of-n (Lin/Con/k/n) system, the reliability calculation method of components and the system with reassignment are derived, respectively. Then, the Delta importance (DI) based heuristics are proposed to perform component reassignment (CR). The heuristics utilize the importance ranking based on DI, searching the most important component to be reassigned iteratively. Third, CR problem (CRP) is presented to find the optimal time and scheme of reassigning components for maximizing the lower boundary of the system reliability during the mission. To address CRP, DI-based greedy algorithms are developed. Finally, the performance of the proposed methods is demonstrated by two experiments. Experiment 1 illustrates the effectiveness and efficiency of DI-based heuristics. Experiment 2 demonstrates the effectiveness and efficiency of DI-based greedy algorithms. In three situations, the average reliability improvement using DI-based genetic algorithm-greedy (DIGA-G) is improved by 24.35%, 19.98%, 12.95%, respectively. The results show that DIGA-G can achieve higher system reliability for CRP compared with the DI-based two-stage-greedy (DITS-G) algorithm.