Optimization of the production scheduling problem with maintenance

Abstract

To improve production efficiency and save cost, production scheduling is usually integrated with maintenance planning. However, in most of the existing research, only the maintenance on machine is considered. On the other hand, mould maintenance is also a vital element in the plastic industry and the breakdowns caused by the mould are even more than the breakdowns caused by the machine. In this connection, it is necessary to integrate mould maintenance into the production scheduling with maintenance problem. Traditionally, a joint scheduling approach is used to deal with the Production Scheduling with Mould Maintenance (PS-MM) problem, which means that the production scheduling, machine maintenance, and mould maintenance are decided jointly. In addition, some algorithms based on Genetic Algorithm (GA) are used to obtain good optimization solutions. Even so, there is still a mismatch among the production scheduling, machine maintenance and mould maintenance in the existing research. Some new strategies and new algorithms need to be explored to provide a satisfactory solution to this problem. The objective of this research is to analyse and optimize the production scheduling problem with mould maintenance. After analysing the character of the original integrated problem deeply, a problem decomposition mechanism is proposed. Based on this mechanism, a new meta-heuristic algorithm is put forward. Particularly, a multi-maintenance strategy that is seldom included in the traditional integrated problem is considered in the new proposed problem. The maintenance strategy depends on the state of the resource and has different effects on the resource. A Three-Level Particle Swarm Optimization (TLPSO) algorithm is used to solve the problem. Furthermore, uncertainty is firstly considered in the integrated problem. The processing time and the maintenance time are represented by triangular fuzzy numbers instead of determinate numbers. Two objectives, i.e. the robustness and the fuzzy makespan, are optimized. This research deepens the study of the production scheduling problem with mould maintenance and can provide a vital reference for future research on the PS-MM problem. Overall, three deliverables are generated in this research. Firstly, a Three-level Particle Swarm Optimization with Variable Neighbourhood Search Algorithm (TLPSO-VNS) is proposed to solve the addressed problem and is proven to outperform other existing algorithms. Secondly, based on the problem decomposition mechanism, a Three-Level PSO (TLPSO) algorithm is designed for the multi-state multi-maintenance integrated problem. Numerical experiments have shown the advantages of the mechanism and the algorithm. Thirdly, a Multi-Objective Pigeon Inspired Optimization (MOPIO) algorithm is proposed for the fuzzy production scheduling problem considering mould maintenance. The efficiency and effectiveness of the mechanism used in the MOPIO have been proved by the comparison algorithms including Non-dominated Sorting Genetic Algorithm II (adapted NSGA-II) and Multi-Objective Particle Swarm Optimization (MOPSO) algorithm. The problem decomposition mechanism and the algorithms proposed in this research are instructive for other integrated problems. The future research will focus on proposing some efficient strategies to make some improvement in the optimization performance of algorithms proposed in this research.