Selecting a set of best stochastic inventory policies measured by opportunity cost

In the present article, we consider the stochastic inventory model aiming at selecting a subset of best policies with minimum total cost. When the computational budget is limited, then the optimal computing budget allocation (OCBA) is used for allocating the available computational budget to the different policies in order to correctly select the best policies. To measure the quality of the selection, we use the expected opportunity cost (EOC) approach that measures the total difference means between the actual best policies and the selected policies. The proposed algorithm OCBA based on EOC is implemented on a stochastic inventory example. The results show that the EOC approaches zero as the total number of computational budgets increases and that using the EOC measure gives better results than using the probability of correct selection (PCS). Moreover, the numerical results indicate that the proposed algorithm is robust.