Research On Multi-level Flow Shop Scheduling And Dispatching For Automobile Mixed-flow Production

In order to meet the individual needs of consumers,automobile manufacturers adopt a “large-scale,customized” mixed-flow production,which not only puts forward higher requirements for the hardware level of flexible production lines,but also brings new challenges to the formulation of production plans and its scheduling strategies.Since the traditional production scheduling model cannot meet the current needs,in order to ensure the efficiency and quality of automobile mixed-flow assembly,while reducing production and inventory costs,based on the Chongqing Science and Technology Commission project "Automobile production line adaptive intelligent management and control system development and application project"(number cstc2019jscx-mbdx X0014),this article takes the multi-level flow shop(assembly,painting,welding,buffer platform)as the research objects,refines the key problems encountered in the mixed-flow production process,establishes mixed production scheduling and dispatching models,designs intelligent solving algorithms and dynamic scheduling schemes,and verify their effectiveness through comparative experiments.The main work of the thesis is as follows:(1)Summarizes the layout and production characteristics of the multi-level flow workshop(welding,painting,final assembly,buffer zone)under the automobile mixed-flow production mode,The main assembly line of the final assembly shop has been refined to have bottlenecks,optional station overload and manual operation station assembly work lag problems.There are problems in the paint shop that frequent switching of the body color leads to reduced production efficiency and irregular cleaning of the spray gun,which affects the quality of the spray paint.Frequent switching of vehicle types in the welding workshop led to inaccurate fixture positioning and affected the quality of the white body.The linear buffer scheduling scheme could not guarantee the effective implementation of the mixed-flow scheduling plan.(2)In order to solve the problems encountered in mixed-flow production in multi-level flow workshops,a mixed-flow scheduling and scheduling model was established.a bi-objective car sequencing model with load balancing at the bottleneck selection as a primary criterion,and minimizing the number of processing delay when considering the replacement and advanced operation time as a secondary criterion is created for assembly workshop.A multi-objective optimization mathematical model is established that takes the linear buffer "the smallest number first" as the constraint condition,equalizes the vehicle outbound,and minimizes the number of vehicle colors and welding fixture types.A dynamic monitoring plan for the buffer and production line is proposed,and a virtual simulation model for dynamic scheduling of buffer reordering is established.(3)Two intelligent algorithms and a dynamic scheduling algorithm for buffer reordering are designed to solve the established mixed flow scheduling and dispatching model.Firstly,an improved ant colony algorithm based on specific heuristic rules and an improved genetic algorithm based on reordering are designed to solve the mixed production scheduling model of the assembly shop,paint shop and welding shop,respectively.Afterwards,an improved genetic algorithm based on reordering is proposed.Before each iteration of the calculation,a reordering operation is added to make the solution sequence automatically meet the buffer constraints,and then the reverse transformation step is added during the running process,and two different selection operations are set.In addition,according to the linear buffer simulation model,the inbound and outbound strategy and the tracking algorithm for offline repaired vehicles are designed.(4)The established mixed-flow scheduling and dispatching model and the corresponding solution algorithm are experimentally verified by using the actual production data of the automobile companies.Through comparative simulation experiments,it is proved that the mixed-flow scheduling model established in this paper,as well as the improved ant colony and genetic algorithm,can solve all levels of flow shop constraints.The buffer dynamic reordering scheduling model can ensure that the mixed-flow scheduling plan is effectively implemented,thereby increasing production effectiveness.