Reliability optimization of linear consecutive-k-out-of-n: F systems driven by reconfigurable importance

Abstract As a typical kind of reconfigurable system, linear consecutive-k-out-of-n: F (Lin/Con/k/n: F) systems have a wide application in some complex engineering systems, such as oil pipeline systems, spacecraft relay station systems, and batch sampling-based quality control systems. Considering the high-reliability requirements for complex systems, how to improve the system reliability cost-effectively is critical to economically grantee the high reliability and long lifetime of practical systems. The reconfigurable features of Lin/Con/k/n: F systems provide a feasible integrated method, combining component rearrangement and component maintenance, to maximize the system's reliability after a period of operation. Moreover, the reconfigurable importance (RI) can evaluate the contribution of a component to improve the system reliability, and RI also can assess the relationship between component reliability and the optimal arrangement. Thus, a reconfigurable importance-based genetic algorithm (RIGA) is proposed to solve the reliability optimization problem of Lin/Con/k/n: F systems by utilizing the advantages of RI and GA. Experiment 1 (for typical systems) and Experiment 2 (for different types of components) are designed to illustrate the cost-effective characteristics of RIGA. In addition, a numerical example of production monitoring systems (PMS) further demonstrates the effectiveness of RI for guiding the solving process of the reliability optimization problem.