Author Type

Graduate Student

Date of Award

2001

Document Type

Thesis

Publication Status

Version of Record

Submission Date

October 2026

Thesis/Dissertation Advisor [Chair]

Mahgoub, Imad

Thesis/Dissertation Co-Chair

Ilyas, Mohammad

Abstract

Existing MANET routing protocols can be classified into three categories: On-demand, Proactive, and Proactive-On-demand hybrid protocols. This dissertation presents a novel hybrid MANET routing protocol, Adaptive On-demand MANET Routing (AOMAR) protocol. AOMAR borrowed the concepts of temporal location updates in operating systems and cache memories to reconstruct broken routes. Each route entry in a mobile host will have a reference field that is set to indicate whether the route is used within a certain period of time called refresh time. Using the refresh time parameter and the route reference field, AOMAR behavior can be controlled to behave as pure On-demand, Adaptive, or history related Proactive protocol. The three versions of AOMAR, On-demand, Adaptive and Proactive, have been simulated using different network sizes, loads and mobility patterns. Three performance metrics---throughput, delay and overhead---are collected. The results show that AOMAR on-demand has the same pattern of results when compared with other exiting on-demand protocols. On-demand AOMAR shows better performance when compared to Adaptive AOMAR for simulation sessions with uniform workloads. On the other hand Adaptive AOMAR outperforms On-demand AOMAR for simulation sessions with non-uniform workloads. Proactive AOMAR reconstructs recently used routes that have broken. Therefore, its routing overhead results change with the change of the network loads and mobility rates. This is not the case for traditional proactive protocols. Proactive AOMAR has lower control overhead and saves the bandwidth when compared with traditional existing proactive routing protocols. In addition to developing and simulating AOMAR, a new mobility model is proposed. Synchronized Random Direction Model of Motion with Constant Speed (SRDMCS) is a model that is suitable to emulate objects that move in random directions with a constant speed and within a certain area of movement. A new formula, Movement Related Update (MRU), is proposed and implemented. In MRU, the refresh period parameter value is linked to the network mobility level. Using MRU in the simulation gives AOMAR a third dimension such that the protocol not only adapts with how frequently the routes are used and how frequently they are broken, but also with how dynamic the network is.

Extent

175 p.

Comments

College of Engineering and Computer Science|FAU Electronic Theses and Dissertations Collection|Thesis (Ph.D.)--Florida Atlantic University, 2001.|College of Engineering and Computer Science

Identifier

FADT11972

Format

pdf

Download

Share

COinS