Determination of Parameter Settings of Ant Colony System for the Traveling Salesman Problem
| Type of publication: | Techreport |
| Citation: | preusse2009studienarbeit |
| Year: | 2009 |
| Month: | July |
| Institution: | Otto-von-Guericke-Universität |
| Address: | Magdeburg |
| Abstract: | Many computational problems are too complex to be practically solved to proven optimality. This is one of the main reasons why heuristics algorithms using problem speci c knowledge to compute near-optimal solutions became very popular. Ant Colony Optimization (ACO) is one of the most successful heuristics for the Traveling Salesman Problem (TSP), which is to nd a shortest tour through a given list of cities. Unfortunately we cannot use the full potential of ACO algorithms if we cannot determine good parameter settings. Commonly, algorithms use a standard parameter setting, which might be good for many problems. Nevertheless, there are problems for which better settings can be found. This is why we dedicated this work to the search for measures helping us to characterize TSP instances to derive appropriate parameter settings for ACO. We have developed new adaptive settings which have been shown experimentally to be better than the standard version. |
| Keywords: | |
| Authors | |
| Added by: | [GR] |
| Total mark: | 0 |
|
Attachments
|
|
|
|
|
Notes
|
|
|
|
|
|
Topics
|
|
