Multi-objective Metaheuristic Algorithms for Precast Production Scheduling Problems

Abstract

Researchers have developed multi-objective precast production scheduling models (MOPPSM) to describe practical constraints and objectives encountered in precast manufacturing. To adapt to realistic customer orders, this study improved MOPPSM by considering lot delivery of precast components (MOPPSM-LD). Furthermore, we firstly proposed two competitive metaheuristics called multi-objective variable neighbourhood search (MOVNS) and non-dominated sorting genetic algorithm II (NSGA-II) to optimise both MOPPSM and MOPPSM-LD. The performance of the two algorithms, measured by spread and distance metrics, were compared with a benchmark algorithm called multi-objective genetic local search (MOGLS). The experimental results showed that MOVNS and NSGA-II can successfully solve both MOPPSM and MOPPSM-LD problems. And the MOVNS outperformed NSGA-II and MOGLS while the NSGA-II was capable of searching as good as the MOGLS.