MIT-Zaragoza Speaker Series: Amir-Behzad Samii

Prof. Amir-Behzad Samii
MIT-Zaragoza International Logistics Program

 

 

 

” Effect of Allocation Mechanisms on Capacity Reservation Decision Making ”

In many occasions, companies wish to provide customer-differentiated service. These may, for example, be motivated by differences in the perceived customer lifetime value or by contractual service level agreements. In the marketing literature, customer segmentation addresses the issues related to the terms of trade that profitably exploit customers’ different buying behavior. In the operations management literature, inventory rationing and capacity reservation are considered as appropriate measures for balancing differences in revenues and shortage costs across distinct groups of customers. These approaches are typically based on the optimistic assumption that a company can assign a customer specific revenue and penalty cost to any order or unit of demand filled or unfulfilled. In practice, however, it is extremely difficult (if not impossible) to accurately estimate the long term monetary implications of not meeting customer demand and corresponding service level requirements.

Revenue Management provides one example for a successful combination of marketing and operations management aspects. It integrates market/customer segmentation and capacity reservation approaches with the goal of maximizing expected profits. In doing so, it typically relies on two important assumptions: (1) demand for low fare classes arrives before demand for high fare classes (“low before high”) and (2) no long-term implications if demand cannot be fulfilled (i.e. no loss of goodwill, reputation, etc.). Both assumptions are relatively realistic in the Revenue Management’s birthplace and commonly applied field, the airline industry. However, in many other industries these two critical conditions are not satisfied.

In our paper we study a single period capacity reservation problem with independent stochastic demand streams from two customer classes. We do not make very restrictive assumptions about the sequence of demand arrival and focus on service level measures rather than cost/revenue associated with each unit of demand. For each of the three commonly used capacity allocation mechanisms (Partitioned Allocation, Standard Nesting, and Theft Nesting), we develop exact expressions for two conflicting performance measures: (1) the service level of the high priority customers (which is enhanced by capacity reservation) and (2) the service level of the low priority customers (which is reduced by capacity reservation). Based on our service level expressions, a decision maker can analyze the (non-linear) tradeoff between the gain in the high priority service level and the loss in the low priority service level. We also provide analytical insights into the effects that different allocation mechanisms have on the two performance criteria and analyze, under which conditions it is beneficial to utilize partitioned allocation, standard nesting or theft nesting. Our analytical insights are illustrated and highlighted through a set of numerical examples. It is shown that under tight capacity conditions standard nesting, and under ample capacity conditions theft nesting should be the preferred allocation mechanism. Although we limit our analysis to a single period capacity reservation problem, we expect that our results can be generalized to a wide range of problem settings in which a decision maker has to ration a perishable resource among different classes of customers.