Comparative Analysis of Four Prominent Ant Colony Optimization Variants: Ant System, Rank-Based Ant System, Max-Min Ant System, and Ant Colony System

TL;DR

This paper compares four Ant Colony Optimization variants on the Traveling Salesman Problem, analyzing their performance across different problem sizes and providing insights for selecting the most suitable algorithm.

Contribution

It offers a comprehensive comparative analysis of four ACO variants, highlighting their strengths and weaknesses for various TSP instance sizes.

Findings

ACS performs best on small TSP instances.

MMAS is most effective on large problems.

ASRank underperforms compared to other variants.

Abstract

This research conducts a comparative analysis of four Ant Colony Optimization (ACO) variants -- Ant System (AS), Rank-Based Ant System (ASRank), Max-Min Ant System (MMAS), and Ant Colony System (ACS) -- for solving the Traveling Salesman Problem (TSP). Our findings demonstrate that algorithm performance is significantly influenced by problem scale and instance type. ACS excels in smaller TSP instances due to its rapid convergence, while PACS proves more adaptable for medium-sized problems. MMAS consistently achieves competitive results across all scales, particularly for larger instances, due to its ability to avoid local optima. ASRank, however, struggles to match the performance of the other algorithms. This research provides insights into the strengths and weaknesses of these ACO variants, guiding the selection of the most suitable algorithm for specific TSP applications.