A New Hybrid Method for Optimizing Vendor-Managed Inventory Model

This research proposes an effective procedure that is considered in exact solution methods category for solving a nonlinear model of single-manufacturer multi-retailer with multi-product supply chain which uses vendor-managed inventory (VMI) policy. The objective is to maximize total profit for the vendor while satisfying retailers’ benefit under the various operational constraints. Since most of the time there are defective goods in each delivery, this model investigates the inspection and screening process in which imperfect quality items are rejected and the retailers may encounter the shortage at the end of each cycle. Due to complexity and nonlinearity of the model, in the first step genetic algorithm (GA) is used to generate a suitable solution. Then barrier algorithm with applying the steepest descent method (GA/barrier) is developed to further optimize the solution acquired by GA and to achieve the more efficient solutions concluding the wholesale and retail prices of items, replenishment cycles and backlogging rates as decision variables. At the end, the efficiency of GA/barrier and GA algorithms are compared together in a numerical example. The results show that GA/barrier algorithm is able to provide significantly more efficient solutions comparing to GA.