Pareto-Improving Coordination Policies in Queueing Systems with Independent Service Agents

Abstract

Coordination policies that balance server loads are well known for improving queueing-system performance. However, in queueing systems with independent service agents that may decline to participate in the coordination mechanism, a sustainable policy must not only benefit the system, but also each individual agent. As a motivating example, we use the emergency-medicine setting: Emergency departments act as independent agents and overcrowding has a direct impact on the service quality. In such settings, agents are interested in performance measures that address the expected number of patients and the risk of being overcrowded; at the same time, the agents do not want to see a reduction in the expected revenue or market share as a result of coordination. We propose coordination policies that provide improvement on multiple performance measures for all agents and guarantee that the expected arrival rate and hence, the expected revenue, is preserved for each agent in the system.

Keywords: Flow control, Queueing control, Pareto Improvement, Stochastic orders