A BI-OBJECTIVE REDUNDANCY ALLOCATION PROBLEM WITH TIME-DEPENDENT FAILURE RATES

This study considers a bi-objective redundancy allocation problem without component mixing for a series-parallel system. In the mathematical model of our solution, it is hypothesized that the redundancy policy is active. As the failure rate of system components depends on the number of working components, decreasing the number of active components increases the failure rate of working components. The structure of the system is series-parallel with k-out-of-n subsystems and time dependent components failure rates. This problem aims to select the component type among available components and the number of allocated ones to each subsystem with weight and volume constraints to maximizing system reliability and minimizing system costs. Since the model is nonlinear and complicated, we suggest the Non-Dominated Sorting Genetic Algorithm II, Non-Dominated Ranking Genetic Algorithm and Multi objective evolutionary based on decomposition to solve the problem. Using the statistical test, is shown this algorithm has better performance.