A Greedy Randomized Adaptive Search Based Two-stage Solution Approach for Scheduling Flexible Manufacturing Systems with Multi-fixturing Pallets

Abstract

This study considers an operations scheduling problem for flexible manufacturing systems with multi-fixturing pallets that can load multiple parts for the objective of minimizing makespan. The problem consists of loading of parts to pallets, i.e. palletizing, input sequencing of pallets to be released initially into the system, i.e. pallet input sequencing, routing of pallets through loading/unloading station and machines, i.e. pallet routing, and sequencing of pallets to process their loaded parts on each machine, i.e. pallet process sequencing. Practical constraints such as central buffer capacity, limited number of pallets and availability of pallet setup workers are considered. For a given set of pallets in the central buffer, the pallet routing/process sequencing sub-problem can be regarded as an open shop scheduling problem. Based on this, a two-stage solution approach is proposed that palletizing/pallet input sequencing are done sequentially using a priority rule and the resulting pallet routing/process sequencing sub-problem is solved using a greedy randomized adaptive search procedure (GRASP) that consists of construction and local search phases. Computational experiments were done on a number of test instances and the results show that the two-stage approach proposed in this study outperforms the existing one significantly.