We consider the problem of dynamic admission control in a Markovian loss queueing system with two classes of customers with different service rates and revenues. We show that under certain conditions, customers of one class, which we call a preferred class, are always admitted to the system. Moreover, the optimal policy is of threshold type, and we establish that the thresholds are monotone under very restrictive conditions. Copyright © 2001 by Marcel Dekker, Inc.

In this paper we study a situation in which a broker must manage the procurement of a short-life-cycle product. As the broker observes demand for the item, she learns about the demand process. However, as is often the case in practice, it becomes either more difficult or more expensive to procure the item as the selling season advances. Thus, the broker must trade off higher procurement costs against the benefit of making ordering decisions with better information about demand. Problems of this type arise, for example, in the travel industry, where a travel agent's cost of procuring airline and hotel reservations increases as the date of a vacation package approaches. We develop a newsvendor-like characterization of the optimal procurement policy. In a numerical analysis, we demonstrate how broker procurements tend to cluster just before price increases and how brokers can benefit from explicitly considering the effects of information about demand in their ordering policies.