Assembly Sequence Optimization Based On Genetic Algorithm

In this thesis, based on the washing machine production line of Haier-Merloni Inc, the genetic algorithms are applied to find the feasible and optimal assembly sequence. The constraints of manufacturing resources are introduced to make the sequence more accurate. According to the information collected from the assembly line, the mathematic model is firstly established, and the optimization criteria such as makespan, accessibility and scheduling cost are concerned. Then the coding scheme and genetic operators are designed through which the static and dynamic optimization is discussed respectively. Because it's a multi-objective and multi-constraint optimization model, the hereditary information is easy to be destroyed by the procedure constraints. Therefore, in the process of optimization, the premature convergence and the fault of effective gene are caused. To resolve this problem, several new genetic operators are proposed, such as gene repairing operator and adaptive operator. Combining with local searching, the hybrid genetic algorithms show great effectiveness. Furthermore, the global convergence of the algorithms is analyzed. At last, the optimal assembly sequence is compared with which adopted in Haier to testify the validity of the proposed approach.