A heuristic for setting up a flexible assembly system

We consider setup problems in flexible assembly systems (FASs) with an objective involving job priorities and station workloads. Because of the difficulty in solving the FAS setup problem, it is partitioned into two subproblems: the part type selection problem and the loading problem. We suggest a solution procedure in which branch-and-bound methods and heuristic methods are used for solving the subproblems. Part types are tentatively selected first without explicitly considering the system's restrictions, such as precedence relationships among tasks. As a result, the alternative for part type selection may not always be feasible for the loading problem. For a feasibility check and for a use in assigning tasks of selected part types, we develop an algorithm for finding a feasible sequence of the tasks. In the loading problem, tasks are assigned to stations using LPT-type rules with improvements tried through interchanging tasks. Results of computational tests on randomly generated problems are reported.