MOEA/D‐based multi‐row facility layout optimisation method with discontinuity perceiving of the Pareto front

Abstract In a multi‐row facility layout problem (MRFLP), facilities are arranged in more than one row under the limited layout area. Considering different layout factors, various extensions of MRFLP have been modelled. However, the orientation of input/output (I/O) point in a facility, as a key factor that plays a direct impact on flow cost, is seldom taken into account. Hence, a novel MRFLP is formulated to minimise both the layout area and flow cost, in which the position of I/O point is employed to calibrate the position of a facility and its orientation determines the flow cost for transporting materials among facilities. To solve the problem, a decomposition‐based multi‐objective evolutionary algorithm with perceiving discontinuity characteristic of Pareto front (PF) is put forward. After judging the discontinuity of PF, the weight vectors are adjusted to enhance the exploitation in continuous regions. To avoid premature convergence, a re‐initialisation mechanism is introduced to help the population jump out of the local optima. Moreover, an improved adaptive epsilon constraint‐handling technique is designed to effectively balance the diversity and feasibility of a population during the evolution. Experimental results of 12 instances expose that the proposed algorithm outperforms other state‐of‐the‐art algorithms significantly.