A multi-attribute utility theory approach to ordering policy for perishable items

Most retailers aim at operating with low cost and low stock-out probability. In this paper, we consider a retailer who manages the inventory of a perishable product using a continuous review (Q,r) policy with these two performance measures. We rely on the multi-attribute utility theory to combine these two measures into a single weighted utility function. We incorporate the decision maker's risk-aversion/risk-seeking attitude. We obtain the optimal policy parameters by maximizing the weighted utility function. We conduct a numerical study to demonstrate the sensitivity of the optimal policy parameters to the risk-aversion/risk-seeking attitude of the decision maker and weights of evaluation measures. Our results reveal that the risk behavior does not have a significant effect on policy parameters while the weights of evaluation measures are influential. We compare our multi-attribute utility model with three well-known alternative (Q,r) policies. In the first one, we study the expected cost as a single performance measure. In the second one, we consider a service level constraint and in the third one, we use CVaR measure. We conclude that our multi-attribute approach has multiple favorable attributes compared to its widely used alternatives.