Planning Remanufacturing Systems by Constrained Ordinal Optimization Method with Feasibility Model

Planning for a complex remanufacturing systems is often an NP-hard problem in terms of computational complexity and simulation is usually the only available but very time-consuming approach in many cases. Ordinal optimization offers an efficient framework for simulation based optimization approaches. In this paper, a new constrained ordinal optimization method is presented for solving remanufacturing planning problems. The scheme of "Horse Race" with Feasibility Modeal (HRFM) is developed to select the set of good enough plans. The rough set method in machine learning and knowledge discovery is applied to generate rules for feasibility determination. This method is compared with the Blind Picking with Feasibility Model (BPFM) method. Numerical testing of a practical remanufacturing system shows that the HRFM method presented in this paper is more efficient to meet the same required alignment probability.