Minimizing task reassignments under balancing multi-product reconfigurable manufacturing lines.

This paper deals with optimization aspects of designing reconfigurable production lines with fixed number of workstations. The studied lines are characterized by their ability to handle multiple products, where switching from one product configuration to another is performed by reassigning certain tasks between workstations. Thus, given a set of products and the order of their arrival, the considered optimization problem consists in designing an admissible line configuration for each product such as the total number of task reassignments is minimized. An admissible line configuration can be viewed as a distribution of a given set of tasks among workstations subject to the corresponding cycle time and precedence constraints. To solve this problem, a mixed-integer linear program (MILP) is proposed. Then, an appropriate MILP-based heuristic, named as HALT-AND-FiX, is developed to efficiently tackle large scale instances. Computational results are reported on a collection of instances derived from the well-known benchmark data used in the literature for the simple assembly line balancing problem. These results show that the heuristic has demonstrated a good overall performance when dealing with large size instances.