Managing Queues with Different Resource Requirements

Abstract

In many service systems, customers from different classes may have very different resource requirements. These differences include not only the duration of the service but also the units of resource required. This is especially prominent in healthcare settings, where patients with different severity levels can require a different level of medical attention/supervision translating into varying demands of nurse staffing capacity. Motivated by these applications, in “Managing Queues with Different Resource Requirements,” Zychlinski, Chan, and Dong study the optimal scheduling policy of a multiserver queue in which customers from different classes may require different units of servers. When customers have heterogeneous resource requirements, in addition to the cost of waiting and resource requirement, we also need to take policy-induced idleness into account. A good policy needs to carefully balance the myopic cost reduction rate and the more forward-looking system throughput. An index-based policy, referred to as the idle-avoid [Formula: see text] rule, is developed to balance these factors. Theoretical analysis along with numerical experiments provide support for good and robust performance of the proposed policy.