Joint decisions on inventory classification, supplier selection and production policy for a multi-item EPQ inventory system under uncertainty

There is a great need to improve the classical inventory models so that they can address real-world problems more properly. The presence of multiple products and a variety of inventory items have complicated the inventory control process, so companies need to classify inventory items to reduce costs. On the other hand, the supplier selection problem is important, as there may be several suppliers with different options in the market. Also, several factors impact the demand for products and cause uncertainty for this parameter. This research develops a multi-product EPQ model that simultaneously classifies products, selects the best possible supplier for each group, and determines the replenishment policy under uncertainty in demand. To solve the proposed model, we present a simulation-optimization approach. This approach uses genetic and simulated annealing metaheuristic algorithms to solve the problem. Also, there is a simulation module that helps the algorithm to evaluate the fitness function. The parameters of algorithms are tuned by employing the Taguchi method. The results are analyzed for three categories of examples. Finally, the sensitivity of the objective function to the input parameters is also analyzed. We found that the system's total cost is highly sensitive to products unit holding cost.