The Utilization Of The Simple Multi Attribute Rating Exploiting Ranks Can Enhance The Performance Of The Aco Algorithm

Abstract

In a comparative study, the performance of the ACO algorithm and a modified genetic algorithm (MGA) were evaluated for solving the multiple salesman traveling problem (MTSP) using various datasets from TSPLIB. The results revealed that although the proposed algorithm did not achieve the best solution, it exhibited improved time efficiency as the dataset size increased. The objective of this study is to improve the performance of the ACO algorithm by integrating the SMARTER algorithm, which aims to find the optimal route and minimize travel time. The combination of these algorithms offers alternative path solutions that can be effectively applied to solve TSP case examples and advance the development of new algorithms that excel in identifying the closest path. The study utilized TSPLIB datasets ranging in size from 76 to 1002 cities, sourced from the Felts and Nelson Krolak repositories. Within this study, the ACO algorithm was employed to optimize the overall distance in the TSP dataset, while the SMARTER algorithm generated suggestions for the optimal routes based on the total trip distance calculated by ACO. Experimental results demonstrated that the ACO algorithm, combined with SMARTER, achieved an average time improvement of 74.09% compared to the MGA algorithm, representing the most optimal performance.