Research On Parallel Machine Scheduling Problem With Preventive Maintenance And Job Release Time

With the implementation and promotion of the Made in China 2025 strategy,the enterprises are gradually transforming and upgrading to digital and intelligent manufacturing,and the production mode has shifted from traditional mass production to mass customization.In order to maintain the stability of enterprise performance under dynamic market demand,it is particularly important to improve the management level of production scheduling.Because the direct factor that affects the efficiency of an enterprise is the utilization rate of the key processes of the enterprise,and these processes present the form of parallel machine production,the Parallel Machine Scheduling Problem(PMSP)has become the focus of enterprise production scheduling.At the same time,production disruption due to machine downtime affects the output and benefits of the enterprise.Therefore,research on PMSP with preventive maintenance is of great practical significance.PMSP has been proved a NP-Hard problem because of its multi-constraint and multi-objective characteristics.The literature lacks an effective intelligent algorithm for PMSP that integrates preventive maintenance and job release time,so the research in this paper has important theoretical significance.This paper addresses the identical PMSP with preventive maintenance and job release time.Mixed integer programming(MIP)models,heuristic algorithms and improved NSGA-II algorithms are designed.They are compared in the operation efficiency and quality.It provides the basis for the enterprise to improve the production scheduling level and improve the efficiency,and lays the foundation for the digital and intelligent transformation of the enterprise.The main research contents are as follows:Firstly,the basic concept of identical PMSP with preventive maintenance and the common solution methods of PMSP are summarized,which provides a basis for determining the choice of solution methods for the research problems in this paper.Secondly,the P2/nr,FPM/Cmax problem with integrated consideration of constraints such as flexible preventive maintenance and nonresumable job is studied,with the goal of minimizing completion time.Two MIP models for this problem are constructed,three heuristic algorithms and improved algorithms based on insert and swap operations are designed.The experimental comparison shows that MIP ? has better operation efficiency and quality than MIP ?.The heuristic algorithm SWA(Sorting While Allocating)and its improvement methods are better than other methods and have the best robustness.Finally,the Pm/rj,nr,FPM/(Cmax,NT)is formed based on the P2/nr,FPM/Cmax problem,considering the constraint on the release time of the job and the objective of minimizing the number of tardy jobs.A mixed integer-programming model for this problem is constructed,and NSGA-? and modified NSGA-? algorithms(M-NSGA-?)are designed.The experimental results of actual enterprise data show that for small-scale problem(n ?12),the effectiveness of MIP is better than M-NSGA-? and NSGA-?,but the efficiency of MIP is poor.For medium-and large-scale problem(n>12),although the diversity of solutions of NSGA-? is better than that of M-NSGA-?,the efficiency and effectiveness are worse than that of M-NSGA-?.