Date of Award

Spring 2001

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Electrical Engineering - (Ph.D.)

Department

Electrical and Computer Engineering

First Advisor

Symeon Papavassiliou

Second Advisor

Ali N. Akansu

Third Advisor

Nirwan Ansari

Fourth Advisor

Richard A. Haddad

Fifth Advisor

Lawrence Ho

Abstract

The proliferation of Internet and demand for wireless services necessitate large increases in capacity and data rates in order to support different multimedia services. Wireless systems will be required to support sources with a variety of traffic characteristics and quality of service requirements. This requires algorithms for admission control and resource allocation at the session, burst and packet levels. The purpose of this research is to develop and analyze optimal resource allocation strategies to maximize throughput of wireless systems with integrated services. Given the multimedia user requirements, the problem addressed can be formulated as a constrained optimization problem. The objective of the admission control and resource allocation policy is to determine the channel allocation to the users with the corresponding feasible power and rate vectors such that overall traffic carried by the system is maximized when all quality of service constraints are met.

The thesis is focused on Code Division Multiple Access (CDMA) wireless networks where transmission modes of the users are controlled according to their QoS requirements and traffic characteristics. The problem under consideration is to find an optimal allocation of traffic channels for a CDMA system with integrated services in order to increase the capacity. Specifically, different channel allocation techniques are examined for data applications, namely circuit, dedicated burst/packet transmission modes and common packet channel transmission schemes. The performance analysis of CDMA common packet channel transmission schemes is studied in depth for finite population and finite buffers/finite sojourn time cases for more realistic data arrival processes than assumed in the literature. The effect of model parameters and user characteristics such as packet length distribution on the system behavior is quantified.

Share

COinS