Research In Automobile Assembly Line Balancing Problem Based On Improved Genetic Algorithm

In current automobile industries, large-scale production mode is widely employed during the leading processes in order to lower production costs while, multiple varieties & small batch production mode is used for the final production processes, assembly, so as to response to customers'demands. Therefore, the production management during the process of assembly comes to be increasingly complicated. Actually, the imbalance of workload resulting from task allocations has influenced production management terribly in decreasing production efficiency, lowering the staff's enthusiam and debasing the product quality. Therefore, in order to optimize assembly line production management, improve production performance and meet the needs of automotive assembly industry, it is necessary to study on the assembly line balance problem and solve them by special efficient and effective algorithm.Taking account of the characteristics of assembly line, this paper analyzes the difficulties and significance of assembly line balancing problems, and discusses three types of assembly line balancing problems and their evaluation criteria. In view of three different assembly line balancing problems, it establishes the mathematical model respectively, and studies on each solution method.Due to the difficulties and low efficiency in solving assembly line balancing problem, especially for large scale automobile assembly line balancing problem, the improved solution framework based on genetic algorithm is proposed in this paper, where both sequence based chromosome coding and station based dynamic partition technologies are implemented to ensure the feasibility of all solutions, and corresponding initialization, selection, crossover and mutation operators are redesigned to assure that the assembly line balancing problem can be solved to near optimal solution. Subsequently, relevant parameters of genetic algorithm have been setup after experiments.Finally, with taking three basic type problems and a real automobile assembly line as examples, the validity and feasibility of the improved genetic based algorithm has been proved. And further experiments on standard benchmark cases demonstrate that the proposed can solve assembly line balancing problem to near optimization.Relevant instances and comparisons have testified completely that the improved genetic basd algorithm is convenient and efficient for the station layout and the task allocation optimization in the assembly line balancing problem. Furthermore, it lays a good foundation for the hybrid assembly line balancing problem, and provides some theoretical guidance for the production management in the enterprises.