Date of Award

Spring 5-11-2012

Degree Type

Honors College Thesis



First Advisor

Dia Ali

Advisor Department



Mobile agents are a relatively new topic in the realm of computer science research. They are being researched throughout the world by many scientists in order to ascertain their viability in real world applications. They have many disadvantages that keep them from widespread adoption. With his graduate dissertation Genetically Engineered Intelligent Mobile Agents, Kackley opened a whole new area of this research by combining mobile agents with genetic algorithms. These algorithms model the natural process of evolution to evolve solutions for problems.

This thesis is part of a research group effort to expand on Kackley’s work in the Database Research Lab for Intelligent Agents at the University of Southern Mississippi. The goals of this research were to first gain a thorough understanding of both mobile agents and genetic algorithms and to augment DNAgents 2.0 with new capabilities. Currently, in Kackleyʼs implementation of DNAgents 2.0, there is no mechanism to facilitate a changing network topology. This behavior is not suited for a live network environment in which computers could be connecting or disconnecting to the network throughout the mobile agentʼs task. In order to take advantage of the full potential of mobile agents, one needs to give them ability to adapt to an ever-changing network. In this research we propose an addition to Kackleyʼs dissertation project to allow adding arbitrary nodes to the network. In our DNAgents 3.0, the simulation is now able to add new nodes with either random or specific links to other nodes. The agents are immediately able to seamlessly move to the new node by reaching it through one of these linked neighbors. This allows for the genetically engineered agents to operate on a dynamic network topology.