Research On Production Sequencing And Execution Management For Mixed-Model Assembly

As a generally adapted production style in current discrete manufacturing industry, mixed-model assembly ranges from differing product, complicated assembly process to variety materials. Its assembly sequencing and executive controlling are, therefore, more complex than the general assembly production. On the basis of experience in the automotive industry, several common denominators in the mixed-model assembly sequencing and executive controlling are generalized. Primarily, sequencing requirements both in a single workshop and in multi-workshops have been discussed in detail. In the second place, several techniques guaranteeing assembly sequence effectively executed also have been studied.In terms of establishing model for optimization, taking the example of automotive industry, the primary constraints and objectives in a single workshop are analyzed. In order to avoid the drawbacks in the previous studies, buffer-based sequencing has been taken into consideration. In the perspective of correlation of workshops and buffering strategy, the common problems in the mixed assembly sequencing and the executive controlling have been classified. Moreover, a comprehensive infrastructure of mixed-model assembly sequencing and execution controlling have been set up.In terms of mixed-model assembly sequencing in a single workshop, there are two universally employed optimal objects: objective based on workload and objective based on parts usage. Analysis in this dissertation reveals that sequencing based on these two objectives may leads to high frequency changeover hence cutting down the productivity. In tackling this problem, a product similarities based objective is proposed, and relatively model and algorithm have been presented as well. Some computational results demonstrate that by employing product similarities objective, the changeover in assembly can be reduced remarkable. However, results also show it is hard to optimizing when just adopts one of the three objectives mentioned above. Considering the characteristics of the problem, a multi-objective sequencing model is proposed. Computational results show that multi-objective sequencing can reduce the frequency of changeovers in the assembly process while meet the object of workload and parts usage.In terms of assembly sequencing in multi-workshop, buffer between workshops has been studies and much emphasize has been put on two types of mixed-model assembly sequencing problems: sequencing under constraints of upper-stream workshop and sequencing based on selectivity banks. The main feature of the former is that sequence in upper-stream workshop must be guaranteed when creates the sequence of down-stream workshop. According to this feature, a cross-over based on homomorphism product is proposed which has the advantage in reduce the repairing operation in genetic algorithm. To solve the latter problem, a composite algorithm is designed in which genetic algorithm is employed in selecting a lane to push a product into and metaheuristic algorithm is adapted in selecting a lane to fetch a product from. The two problems mentioned above are suit for two kind of assembly process in which FIFO buffer and selectivity bank is employed between two successive workshops respectively. The algorithms in solving both problems adopt delay-binding strategy which is illustrated to have the ability of promoting flexibility of sequencing and resulting in better solution.An excellent sequencing is not enough for accomplishing the production plan, execution management is compulsory also which makes the assembly process an integrated controllable system. It makes the manufacturing system capable of dealing with variety abnormal events with great ease. The closed loops in assembly process includes: sequencing-control loop, material delivery Andon loop and information inquiring-feedback loop. To realize these loops, some studies are focused including online data-acquirement, management of work-in-process, visualize of production data, material match and delivery, integrating with other information systems.Research results in this dissertation have been put into the practice of MES in Jianghuai Automobile Co-operation (JAC). On one hand, the model and algorithm discussed above are applied into actual production in which constitutes sequencing based on product similarities and correlating sequencing in multi-workshops based on selectivity bank. Some improvements are made on algorithms in accordance with specific problem in the enterprise. On the other hand, in order to meet the demand of automotive industry, an automobile assembly MES system is developed which realize the function execution function in body-shop, paint-shop and trim-shop. Practices show that integration of sequencing and control can make the mixed-model assembly a visualable, controllable, robust process. The MES mentioned above has been put into use in JAC for more than six months.