Pruning Algorithm for Multi-Objective Optimization with Decision Maker's Preferences of System Redundancy Allocation Problem

A preference-based ranking method is proposed to identify appropriate solutions among all trade-off solutions. This method expresses the decision maker's preference as a ranking priority of the objective functions. After obtaining the approximation of optimal solutions from a searching algorithm, the proposed pruning approach can specify the promising solutions while satisfying the decision maker's preferences which are specified as raking/relative priority of the objective functions. The proposed pruning method is demonstrated with a redundancy allocation problem of series-parallel system. The results show that the new method is generally effective to identify the preferred solutions. In addition, the proposed pruning method is easy to use even when the decision maker is not an expert in the problem solving.