LoboVault Home

A more robust ant colony learning algorithm : with application to travelling salesman problem

LoboVault

Please use this identifier to cite or link to this item: http://hdl.handle.net/1928/11110

A more robust ant colony learning algorithm : with application to travelling salesman problem

Show full item record

Title: A more robust ant colony learning algorithm : with application to travelling salesman problem
Author: Nandina, Viswanath
Advisor(s): Heileman, Gregory L.
Committee Member(s): Verzi, Stephen J.
Ghani, Nasir
Department: University of New Mexico. Dept. of Electrical and Computer Engineering
Subject: Ant System
Ant Colony optimization
Travelling salesman problem
Non fully connected graphs
TSPLIB
shortest path
pheromone updating
Priority based evaporation
control parameters
Gallants Approach
Elitist Strategy
Average Branching Factor
Dynamically changing nodes
LC Subject(s): Ant algorithms.
Traveling-salesman problem--Computer simulation.
Degree Level: Masters
Abstract: Graph problems model many real life applications, where the quantity of the nodes often changes with time. In such graphs, the evaluation of shortest tour is important as various guiding and navigation systems use this information. Nodes of a graph, in many applications, often change over time, and evaluation of shortest tour is essential whenever a new node is added or deleted. We propose an algorithm that deals with such situations. We have used the Ant System with a different meta-heuristics, to find the shortest tour in a graph. We have analyzed the performance of our proposed algorithm with other algorithms by using the problem instances given in TSPLIB. The proposed modification to the Ant System heuristics will also work for directed and non-fully connected graphs. We show the use of meta-heuristics that make our algorithm free from stagnation, that is, we prevent the ants from taking up the same tour repeatedly which helps to continuously search for better results. Our approach further adopts a method that is a modification to Gallant’s Technique, to choose the appropriate convergence within the reasonable computation time.
Graduation Date: July 2010
URI: http://hdl.handle.net/1928/11110


Files in this item

Files Size Format View
Viswanath_Thesis.pdf 2.281Mb PDF View/Open

This item appears in the following Collection(s)

Show full item record

UNM Libraries

Search LoboVault


Browse

My Account