Original scientific paper

https://doi.org/10.31803/tg-20230825125127

Reducing ACO Population Size to Increase Computational Speed

Luka Olivari orcid.org/0000-0003-2214-1552 ; Šibenik University of Applied Sciences, Trg Andrije Hebranga 11, 22 000 Šibenik, Croatia *

* Corresponding author.

Abstract

Ant Colony Optimization (ACO) is a powerful metaheuristic algorithm widely used to solve complex optimization problems in production and logistics. This paper presents a methodology for enhancing the ACO performance when applied to Traveling Salesman Problems (TSP). By reducing the number of ants in the colony, the algorithm's computational speed improves but solution quality is sacrificed. An optimal number of ants to produce the best results in the shortest time is specific to the problem at hand and can't be defined generally. This paper investigates the effect of ant population reduction relative to the problem size by measuring its impact on solution quality and execution time. Results show that for certain problem sizes ant population and execution time can be halved with practically no reduction in solution quality, or they can be reduced 5 times at the price of slightly worse solution quality. Reduction of ant population is much more impactful on reduction of execution time than it is on solution quality.

Keywords

ACO; Ant Colony Optimization; colony population; number of ants; speed up

Hrčak ID:

321393

URI

https://hrcak.srce.hr/321393

Publication date:

5.12.2024.

Visits: 113 *