Research On Mixed Model Two-sided Assembly Line Type Ⅰ Balancing Problem Based On AM Genetic Algorithm

In order to satisfy customers’various demands and improve company s’competence,the mixed model assembly is applied.Meanwhile,the style of assembly line appeared variable,from single-sided assembly line to two-sided assembly line,U-type assembly line and so on.Among them,the two-sided assembly line is used to assemble the large-size products,such as cars,loaders,etc.The two-sided assembly line has several advantages over traditional single-sided assembly line,including shortening the line length,enhancing the usage rate of devices and improving assembly workers’ production rate and so on.Accompanied with the benefits,there are several new constraints when balancing two-sided assembly line.The operational direction of task,the sequence-related finished time constraints,and so on,should be taken into account,besides the common constraints when balancing single-sided assembly line.Hence,the mixed model two-sided assembly line balancing problem becomes more complex and needs to develop new balancing technique.According to this problem,the mixed model two-sided assembly line balancing problem is studied from following aspects in the thesis,including the mathematical model,the methods and the real project application.The main researches are as follows:(1)Build the mathematical model of mixed model two-sided assembly line balancing problemThe mathematical model is built which is based on the satisfaction of two-sided assembly line’constraints.The three minimal objectives are the station number,the balance index and the wait time.The model makes the mixed model two-sided assembly line balancing problem solved with mathematical method possible.(2)Formulate AM genetic algorithm(AMGA)① A real number combination based on task and operational direction is adopted in the thesis.In the encoding,the length of chromosome is equal to twice the size of the problem to be solved,odd bits of genetic of the chromosome are the value of assembly tasks,even bits are the value of direction which the tasks are assigned to.② A dynamic mutation rate is used.With the improvement of evolution algebra,genes of the chromosome are more and more excellent,the mutation rate is becoming smaller and smaller,since it not only reduces the variation rate of the population but also avoids the disorder of the populations’ evolution and makes the result approach from better to the optimal.③ In order to maintain the diversity of the population and avoid slow convergence of parallel selection method,the combination of the parallel selection method and arrangement selection method is used in the thesis,after some algebras,pareto optimal solution can be obtained.(3)With the experiments of two-sided assembly line balancing problem which are pulished,the results show that the advanced multi-objective genetic algorithm outperformed other heuristic methods.(4)A practical assembly line balancing problem of Ycompany is also solved by this AMGA.If keeping the original cycle time(120s)constantly,a solution with 8 stations,26.42509s balance index and 22.83383s waited time less than original scheme can be found.So,a new,fast and effective solution is provided to solve the mixed model two-sided assembly line balancing problem of type I in the thesis.And the solution is creative and theoretical value.