Ant Colony Algorithm For Two-sided Assembly Line Balancing Problem

Two-sided assembly lines are widely used in large and complex mechanical products’ assembly lines. The two-sided assembly line can shorten the length of the assembly line and at the same time increase the efficiency of workers and fixtures compared to the sample assembly line. As the special structure of the two-sided assembly line, such as the restriction of each task’s operation directions and operating constraints of the sequence of tasks, it is more complexity to design a two-sided assembly line than a sample assembly line.This article analysis the status quo of two-sided assembly line, then find out the researching defects in today’s research work and formulate the research field. After study the characteristics of the two-sided assembly line, proposed a new mathematical model. Then make a in-depth study in the following problems:two-sided assembly line balancing problem of type-Ⅰ, stochastic two-sided assembly line balancing problem of type-Ⅰ, two-sided assembly line balancing problem of type-Ⅱ, two-sided assembly line balancing problem of type-Ⅰ with zoning constraints.This article first analysis analysis three basic heuristic algorithms which are based on only one different heuristic rule for two-sided assembly line balancing problem of type-Ⅰ in order to select the excellent heuristic rules for reference in the future work. After this work, proposed two hybird heuristic algorithms for stochastic two-sided assembly line balancing problem of type-Ⅰ, then analysis the performance of these two algorithms.Then proposed an ant colony algorithm that formulates the task sequence before generates an answer by using the task assignment rule. After that research, developed another ant colony based algorithm to solve two-sided assembly line balancing problem with zoning constraints. In this study, an improved ant-colony-based heuristic algorithm is proposed for solving this problem. A improved hybrid ant-based search rule and a heuristic task distribution rule were used in order to establish a feasible solution which is also in accordance with the zoning constraints.This algorithm has better performance compared with varies algorithms proposed by other researchers. And the Gantt chart display module was added in order to enhance its’practicability.Then, this paper proposed an ant colony based algorithm for two-sided assembly line balancing problem of type-Ⅱ, this algorithm is formulated in the foundation work of earlier period research in this article and also contained other researcher’s viewpoint which plays influential role in finding the optimization cycle time. The algorithm is coded into a program by MATLAB and the usablity is indicated by a practical example.Finally, this paper involves a real two-sided assembly line in one of the largest car manufacturers in China. By using the algorithms proposed in this article, generating several improving proposals and the usablity are indicated.This research is supported by National University Specialized Research Fund for the Doctoral Research Topics(200806131014).