Research On Assembly Line Balancing Problems Based On Simulation Optimization

The appearance of assembly line about a hundred years ago, has largely promote the process of industrializing of the world. And the research of Assembly Line Balancing (ALB) started at about several decades ago, the target of ALB research evolved from the simple models to the multi-product and mixed flow assembly line. The methods used in the research were mostly the mathematical modeling, especially the heuristic approach. But the factors considered in the Assembly Line Balancing, such as failure rate of machines and randomness factors, make the optimization methods of mathematical modeling cannot fulfill the needs of the manufacture in the real world. Computer simulation can add the settings of real world into the simulation models, which only can actualize several simulation experiments, but cannot find the best solution through iterating the simulation model. Simulation based optimization is a new method that optimization method of which the real-valued function is replaced by the simulation model. The process of simulation based optimization is followed: first, the optimization method generates the initial solution, and then input the initial solutions into the simulation model, which generates the results after running the given times; the optimization method will evaluate the results. Thus simulation based optimization can achieve the advantages of both simulation methods and optimization approach that could make the research with realistic significance, also can choose the best solution. This thesis is based on simulation based optimization and with the help of Plant Simulation9.0, actualizes balancing the mixed flow assembly line with six stations. The software Plant Simulation9.0 has the great power on dealing with the data and table; the research defines the process of simulation, inputs the parameters and adjust the policies of controlling through the built-in programming language Simtalk. That could achieve the destination of balancing the assembly line. The innovative points are followed: using the particle swarm optimization algorithm with mutation based on similarity (PSOMS) in the software Plant Simulation9.0, actualize simulation based optimization and then the four algorithms were used in the thesis, the results generated by the simulation experiments were analyzed with the approach of statistics, at the end of the thesis, conclusion that the PSOMS is the best algorithm for solving the problems of assembly line balancing had been made.