Simulation Cutting Approach: Joint Workstation, Workload and Buffer Allocation Problem

The Joint Workload, Workstation and Buffer Allocation Problem (JWWBAP) of open flow lines has been recently investigated in the context of manufacturing system design. This work proposes a simulation cutting approach to exactly and efficiently solve the JWWBAP based on a single sample path. Specifically, a sample path based simula- tion optimization model of the JWWBAP is formulated into a Mathematical Programming Representation (MPR). The MPR is a large mixed integer programming model with high complexity, and a method to approximately solve it has been the subject of our previous research. In this paper, its exact solution is addressed and to this purpose, the Benders decomposition is adopted. The subproblem of the Benders decomposition is an MRP of the simulation aspect of the overall problem and its dual is a network flow problem. We show that both the subproblem and the dual subproblem can be solved using the simulation trajectory. From the solution of the subproblem, a cut is gener- ated and added to the master problem of the Benders decomposition, which corresponds to the optimization aspect of the overall problem. Numerical experiments are conducted to analyze the JWWBAP and the performance of the proposed approach.