earticle

영어

In this study, a perishable inventory allocation model under single-period and two-period settings is considered. The studied model for perishable inventory allocation is a kind of yield management model for firm’s productivity improvement. Yield management is a method for managing capacity or inventory profitably, which was originated from the deregulation of the US airline industry. In the airline industry, seat inventory allocation problem was mostly studied. In this research, we extend the seat inventory problem to a general perishable inventory allocation problem including customer movement. Specifically, we develop a two-price class inventory allocation problem considering customer movement, where the products are sold to different demand classes at different prices. Initially, the number of perishable products is fixed, and the number of sellable products is random at the beginning of next period. For a given number of products at the beginning, the objective is to decide the optimal number of products to assign different classes of customers in period 1 for maximizing the total expected profits. The customer demands are stochastic and independent initially, but dependencies are created by the return of a fraction of customers who could not buy a low-price product in each period. We develop a one-period model and a two-period optimization analytical model with marginal analysis. For the two-period model, we also consider customer movement from period 1 to period 2, where a portion of the customers who could not buy a low-price product in period 1 may wait for reopening of the same low-price item in period 2. The main objective of this study is to develop analytical models to obtain the optimal inventory allocation to support decision-makers in these environments. Though there are many analytical optimization models, it is very difficult to obtain the optimal solution within a reasonable time and effort for practical computation. We also develop a more efficient computer algorithm using marginal analysis techniques. Therefore, we use a marginal analysis in this study. We finally show computational experiments for illustrating the algorithm.