Solving assembly line balancing problem with positional constraints and worker assignments using mathematical programming and heuristic solution approaches

There are many studies in the literature on assembly line balancing (ALB) problems. The ALB problems differ from each other in various aspects. In this thesis, we consider the ALB problem with hierarchical worker assignment, positional constraints, station paralleling options, and task assignment restrictions. The objective of this ALB problem is to decide on the number of parallel stations to be opened in each work stage and to assign tasks and workers to stations such that the sum of station opening costs and worker costs is minimized. To solve this problem, we initially propose an integer programming (IP) model, and then develop a simulated annealing (SA) algorithm to obtain high-quality solutions in reasonable computational times. For the SA algorithm firstly, a modified version of the Rank Positional Weight heuristic is developed to generate an initial solution. To generate new different solutions from the current solution, four kinds of neighborhood search structures are used which are single_transfer, two_transfer, swap and stage separation. In order to enhance the solution quality of SA algorithm, it is hybridized with a local search. We use a giant leap procedure to investigate an inferior or unvisited search space for the probability of finding a better solution. To find the optimal parameters of the SA algorithm, we employ the Taguchi method. Thus, the solution quality and the running time of the SA algorithm improve. A set of test problems are solved using both the proposed IP model and SA algorithm. The computational results show the effectiveness of SA algorithm.