Stochastic Modeling Applications to Service Systems

Date, Time, Venue

January 9th, 2012 at 15.15

Executive Suit, 1st floor, Engineering Front Building University College London

Abstract

We present two applications of stochastic modeling techniques to service systems.

First, we investigate alternative ways to predict, in real time, the delay of an arriving customer in a service system such as a hospital emergency department (ED) or a customer contact center. We envision our delay predictions being used to make delay announcements to arriving customers. Our delay predictors differ in the type and amount of information that they use about the system. We introduce predictors that effectively cope with real-life phenomena, such as customer abandonment (impatience), time-varying arrival rates, and general service-time distributions.

Second, we study a drug dosage adjustment problem. Warfarin is a widely prescribed anticoagulant drug to patients in risk of stroke. However, warfarin has a complex dose-response relationship that makes effective use a challenge. In practice, doctors rely on intuition to prescribe an initial warfarin dose, and then make appropriate dose adjustments until a desired patient response is observed. We propose a new and systematic dose adjustment scheme using a Markov decision process (MDP) model. We validate our model by conducting an empirical study of real-life patient data provided by a hematology clinic in Montreal.

Biography

Rouba is a postdoctoral fellow at the Desautels faculty of management of McGill university. She has a Ph.D. degree in Operations Research from Columbia University (2010) supervised by Ward Whitt. Her main interests are stochastic modeling, simulation, queuing science, call centers, and healthcare operations.