Research On Multi-objective Automated Warehouse Scheduling Method Based On Decomposition

As a new type of warehousing technology,the automated warehouse has its advantages of high space utilization,low labor cost and high access efficiency,and has been received wide research attention.For an automated warehouse with determined hardware configuration,the efficiency of the scheduling process determines the efficiency of the warehouse.With the development of technology and the demand change of enterprises,decision makers hope it can provide different schedules to make decision according to the needs of different scenarios.Therefore,the automated warehouse task scheduling is selected as a main optimizer in this thesis,the multi-objective evolutionary algorithm based on decomposition(MOEA/D)is studied and extended in tihis thesis,and it is successfully applied to solve two kinds of multi-objective automated warehouse scheduling problems.The main topics of this thesis are as follows:1.Considering the characteristics of the automated warehouse scheduling problem,such as order based,this thesis aims to optimize the objective by minimizing both the maximum completion time and the maximum order tardiness,and constructs a multi-objective automated warehouse order scheduling model.An improved MOEA/D which combined the shuffled frog leaping algorithm(SFLA)was proposed to deal with this problem.In order to ensure the steady-state updating of this method,a neighborhood updating strategy based on distance is introduced is this thesis,which is applied to determine which subproblems are need to be updated.A modified SFLA is introduced in the later stage of the method to enhance its search efficiency,and the search effect of the method is significantly improved.2.The enterprise always takes minimum cost to obtain the maximum profit as the operational goal.With frequent inbound and outbound operations in automated warehouse,the equipment energy consumption as the operating cost directly affects the profit of the enterprise.In this thesis,we study the scheduling problem of multi-objective automated warehouse in order to minimize both the maximum completion time and energy consumptions of the stackers.A modified weight vector adaptive MOEA/D is proposed to solve this problem.By introducing a global indicator to control the switching time of spring generation strategy,local search strategy and the population diversity maintenance strategy in this method.The hybrid spring generation strategy can retain the characteristics of different search mechanisms,the local strategy is used to help the method to exhaustively search the objective space and the population diversity maintenance strategy aim to improve the population diversity during the search procedure.Simulation experiments of different problem scales show that the extended MOEA/D studied in this thesis is effective in solving the automated warehouse scheduling problem and hence it is ready to apply to deal with real automated warehouse scheduling problem.