Integrated Optimization Of Production Scheduling And Maintenance Of Discrete Manufactruing Systems

Production scheduling and maintenance planning play extremely important roles in the operational management of discrete manufacturing systems.They interrelate with each other in the system’s availability and reliability,which can be viewed as a couple.Thus,the coordination of them contributes a lot to the efficiency of plant.Based on the present studies in this area,this research focuses on the classic discrete manufacturing systems and proposes the two-dimension integrated model cooperating the production scheduling and maintenance policy simultaneously in order to maximize the total profit of enterprise by providing the scientific decision support to the managers.In the deterministic case ignoring the random machine failures,we focus on the impact of system’s unavailability on the production,which is caused by the performing of preventive maintenances.As the pilot study of flow shops integrated problem,the integration between production scheduling and maintenance planning in the single machine system is studied firstly.The effective working time and the useless idle time of machine are compared.Considering the remaining life of machine,the earliest release date rule and the longest processing time rule are combined to reduce the idle times.Then,the dynamic dispatching rule and the lower bound of problem are analyzed.Based on the analysis,a branch-and-bound algorithm is developed to solve the small-to-medium-sized problems and a fast constructive heuristic is devised to solve the large-sized problems.For the flow shops problem,the finish time of job in the upstream machine can be viewed as the arriving time of job in the downstream machine.Based on the relationship between the jobs and unavailable intervals,the set of maintenance constrants are derived.Then,the mixed integer linear programming mathematical model is built for this deterministic problem.Finally,a hybrid increment genetic algorithm is devised,where the key decision variables are coded into the solution chromosome and the others are decoded heuristically.For each priority list of jobs,the jobs’ sequence in each machine is derived sequentially from the first stage to the last one according to the analysis of single machine problem.The numerical experiments validate the stability and quality of algorithm as well as the necessity and efficiency of the integration model.In the uncertain case considering the unexpected machine failures,the failure function of machine and the probability distribution of machine breakdowns are analyzed firstly.The feature of uncertain factor and the set of possible reaction policies are studied based on which the proactive optimization framework combining the initial plan and the rescheduling scheme is proposed.And,the mathematical models are built for different kinds of manufacturing systems.The production sequence,the preventive maintenance positions and the buffer times are decided simultaneously in order to optimize the bi-objective of quality robustness and solution robustness.For reaction policy I,it is proved that the sub-problem can be solved in polynomial time using the newsboy model.For reaction policy II,the Monte Carlo simulation procedure and efficient surrogate measure are proposed to quickly evaluate the feasible solutions.Then,the iterated heuristic is devised based on the separation of decision variables.The Pareto front is obtained by changing the weight of two objectives,and the balance and compromise between them are also analyzed.The numerical experiments show that the system’s stability is hugely improved by the integration model compared with the traditional mothed,which can guarantee the feasibility and performance of the planning in the uncertain environment.