Semester

Fall

Date of Graduation

2013

Document Type

Thesis

Degree Type

MS

College

Statler College of Engineering and Mineral Resources

Department

Lane Department of Computer Science and Electrical Engineering

Committee Chair

Bojan Cukic

Committee Co-Chair

Donald Adjeroh

Committee Member

James D. Mooney

Abstract

An adaptive system differs from a non-adaptive system in that an adaptive system uses a specific process to identify and implement adaptations to system parameters during run time in an effort to increase system performance.;In order to develop an adaptive system one of the most important aspects is the ability to accurately predict and manage system behavior. If an unexpected event has occurred, accurate prediction of system behavior is needed in order to determine whether or not the system is able to continually meet expectations and/or requirements. When considering any possible adaptations to the system, one must be able to accurately predict their consequences as well.;In a feedback loop of an adaptive system performance analysis and prediction performance analysis and prediction may lead to a large number of different states. Therefore, the method of analyzing the system must be fast. For complex systems, however, the most popular method for predicting performance is simulation. Simulations, depending on the size of the system, are known for being slow.;In this thesis we develop a fast method for prediction of performance of a complex system. We use this method to allocate resources to the system initially, and then make decisions for system adaptation during runtime. Finally we test various modifications to the system in order to measure the robustness of our method.

Download

Share

COinS