Fuzzy Multi-Objective Reliability Optimization of the Mixed Series–Parallel System Using Hybrid NSGA-II

Practically, reliability involved in the system design is enhanced with a reduction of other mutually conflicting objectives simultaneously such as cost, weight, volume, etc. Moreover, different types of uncertainty such as expert’s information character, qualitative statements, vagueness, incompleteness, and unclear system boundaries are typical in multi-objective decision-making of reliability. This paper proposes a hybrid NSGA-II for fuzzy multi-objective reliability optimization which comprises NSGA-II, local search strategy, and clustering technique. Local search strategy helps to update each Pareto-optimal solution after an NSGA-II simulation run which gives a better convergence near the true Pareto-optimal front while the clustering technique maintains a good diversity in the solutions set. Finally, the best compromise solution is achieved by using the fuzzy ranking method. The results obtained by the proposed methodology are then compared with popular elitist multi-objective evolutionary algorithms (MOEAs) namely NSGA-II and PESA-II as well as a multi-objective swarm technique known as MOPSO. A numerical example of the mixed series–parallel is given to show the effectiveness of the proposed approach.