A CRANE SCHEDULING PROBLEM IN A COMPUTER-INTEGRATED MANUFACTURING ENVIRONMENT

This paper addresses a crane scheduling and machine layout problem in a Computer Integrated Manufacturing (CIM) Environment. A single crame is used to move all the Work-in-Process (WIP) in the system. The overall system objective is to maximize the yield rate subject to the flow time limit of the WIP. We formalize the problem, and analytically and empirically show that cyclic scheduling provides a near optimal solution, which is superior to dispatching rules. First, we illustrate the optimality and benefits of cyclic scheduling in a simple environment. Then, for multiple-product problems, we show that for a given sequence, finding the minimum cycle time becomes the maximum cost circular network flow problem in a graph. Based on the insights developed, a heuristic for sequencing product types in a cycle is derived that approximately minimizes the cycle time over all sequences. Finally, computational experiments are reported and various assertions made in the paper are empirically verified.