A Preference-based Multi-objective Evolutionary Algorithm for Redundancy Allocation Problem

Allowing the decision maker (DM) incorporates his/her preferences in Multiobjective Evolutionary Algorithm (MOEA) is likely yield better approximation of optimal trade-off solutions for multi-objective optimization problem (MOP). In this paper, we propose a preference-based MOEA to help the DM choosing the final best solution(s) based on his/her preferred objective(s). Our algorithm is called ASA-NSGA-II. The approach is accomplished by replacing the crowding estimator technique in NSGA-II algorithm by applying an extended angle-based dominance technique. We consider Redundancy Allocation Problem (RAP) to observe the usefulness of our algorithm in practical context. The system composes of multiple subsystems connected in series. The designed objective is to select multiple components for a subsystem. The objective functions are to maximize system reliability (R), minimize system cost (C), and minimize system weight (W), simultaneously. We demonstrate seven cases of preferences with interesting results.