A Genetic Algorithm for solving two-sided assembly line balancing problems

Abstract

Two-sided assembly lines are common industrial practice in the assembly of large-sized products. In this paper a Genetic Algorithm (GA) is developed to solve the two-sided assembly line balancing problem. The developed GA specifies a new method for generating the initial population. It applies a hybrid crossover and a modified scramble mutation operators. A proposed station oriented procedure is adopted for assigning tasks to mated-stations. It specifies the side of the Either tasks based on proposed side assignment rules rather than assigning them randomly. These rules are effective especially in large problems. The proposed method of generating the initial population is able to generate feasible solution in different areas of the search space. The applied genetic operators are able to preserve the feasibility of all solutions throughout all the developed generations. The proposed GA is able to find the optimum and near optimum solutions within a limited number of iterations.