Abstract
This dissertation deals with partitioning a given tool set to different machining centers of an FMS and assigning jobs to those centers in such a way that tool traffic between machining centers is minimized. The number of tool types required for the shop operations is so large for a single machining center that tools need to be partitioned among different machining centers. Each job visits only one machining center for the completion of all required operations, and any unavailable tools are borrowed from other machining centers, which causes tool traffic inside the shop. The tool loading and job assignment problem is formulated as an optimization problem to minimize the number of tool movements between machining centers while maintaining an acceptable workload balance. The problem is decomposed into two subproblems relating to tool assignment and job assignment, and solved iteratively by passing the output of one subproblem as input to the other subproblem. The issues of real time tool management and job dispatching rule are then analyzed using simulation. The solution of tool loading and job assignment through the optimization model is based on the static and aggregate view that the solution may not provide a real time control rule. Here, the effect of tool partitioning and job assignment methods on system throughput is considered together with different job queuing methods and tool return policies of borrowed tools in real time. The problem of determining the optimum number of tool copies for each tool type is considered when the total number of tool copies is limited. The problem is formulated as an optimization model the solution of which provides not only tool copies but also tool assignment and job assignment. For each of the above three related problems, extensive computational experiments were conducted to evaluate the performance of the proposed methods.
Na, Yoon Kyoon (1987). Tool loading and control in a flexible manufacturing cell. Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -746722.