|
Abstract
Ant algorithms were first written about in 1991 and since then they have been applied to
many problems with great success. During these years the algorithms themselves have
been modified for improved performance and also been influenced by research in other
fields. Since the earliest Ant algorithms, heuristics and local search have been the
primary knowledge sources.
This talk will outline an implementation of the Graph-based Ant System and show that
this theoretical model can be used to represent other algorithms such as Ant System and
Max-Min. The talk will then focus on how Ant algorithms make use of heuristics and
local search to find approximate solutions. I will show that Ant algorithms are
ill equipped to deal with misleading heuristic information and that local search
significantly alters the way Ant algorithms behave, resulting in inefficient and
unpredicatable performance when combined in hybrid solutions.
* to get to the Conference Suite, enter at main door at 2 Buccleuch Place from the street, and climb the stairs to the first floor. Enter through the door on the right towards the general office. Go down the internal stairs next to the general office, turn left down the long corridor on the ground floor until you reach an entrance hall. Go down the stairs to your left to the basement and follow the signs to the conference suite.
|
Informatics Forum, 10 Crichton Street, Edinburgh, EH8 9AB, Scotland, UK
Tel: +44 131 651 5661, Fax: +44 131 651 1426, E-mail: school-office@inf.ed.ac.uk Please contact our webadmin with any comments or corrections. Logging and Cookies Unless explicitly stated otherwise, all material is copyright © The University of Edinburgh |
