Reliability-redundancy Allocation Problem with Homogeneous and Heterogeneous Redundant Subsystems

Reliability-redundancy allocation problems (RRAPs) have been widely studied to improve system reliability. In a system, there may be a case where components are homogeneous in one subsystem and heterogeneous in another. However, previous RRAPs only considered the cases of all homogeneous redundant subsystems or all heterogeneous redundant subsystems. In this study, a mixed form of homogeneous and heterogeneous redundant subsystems is considered, and the influence of the increase of the number of heterogeneous redundant subsystems on system reliability is analyzed. Moreover, an improved multi-population genetic algorithm (IMGA) is designed to solve the RRAP. The IMGA controls the spread of advantageous genes among populations through a specific network structure. Experimental results show that the system reliability will increase with the increase of the number of heterogeneous redundant subsystems, and the IMGA with Erdos-Renyi (ER) networks can get better solutions. The comparison with previous studies also proves the superiority of the algorithm designed in this paper.