Date of Award

Fall 2011

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Mathematics, Statistics and Computer Science

First Advisor

Praveen Madiraju

Second Advisor

Sheikh I. Ahamed

Third Advisor

Karl Byleen

Abstract

Unwanted and untimely interruptions have been a major cause in the loss of productivity in recent years. It has been found that they are mostly detrimental to the immediate task at hand. Multiple approaches have been proposed to address the problem of interruption by calculating cost of it. The Cost Of Interruption (COI) gives a measure of the probabilistic value of harmfulness of an inopportune interruption. Bayesian Inference stands as the premier model so far to calculate this COI. However, Bayesian-based models suffer from not being able to model context accurately in situations where a priori, conditional probabilities and uncertainties exist while utilizing context information. Hence, this thesis introduces the Dempster-Shafer Theory of Evidence to model COI. Along the way, it identifies specific contexts that are necessary to take into account. Simulation results and performance evaluation suggest that this is a very good approach to decision making. The thesis also discusses an illustrative example of a mobile interruption management application where the Dempster-Shafer theory is used to get a better measurement of whether or not to interrupt.