An Effective Route Tracing Using Enhanced Ant Colony Optimization Techniques: Travelling Salesman Problem

Abstract

The Travelling Salesman Problem (TSP) is a well-known optimization that determines the shortest route that visits a particular set of towns and returns to the start line. As an NP-hard, its complexity will increase considerably as the number of cities increases. Several heuristic algorithms, such as Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), and Genetic Algorithm (GA), have been developed to tackle this problem efficiently. This work establishes an approach that employs a multi-goal optimization algorithm that optimizes for minimizing the tour distances and maximizing the diversity of the towns visited. This paper aims to apply an enhanced ACO with conventional GA and PSO algorithms separately to leverage the strengths of both algorithms. The PSO and GA paths enhance the pheromone in global exploration, which speeds up the pheromone initialization of ACO. Thus, optimized routes by ant colony get integrated with pheromone reinforcement to obtain enhanced pheromone in the hybrid of GAACO and PSO-ACO. The ACO algorithm applies enhanced state transition to progress the search using the angle guidance function, and the best pheromone level improves convergence speed of the enhanced algorithms. This project has been designed as an intuitive and user-friendly interface for users to input TSP instances, select optimization algorithms, and visualize the solutions generated by each algorithm. Experiments are done to assess overall performance of ACO, GA-ACO, and PSO-ACO based on the best solution and convergence speed. Lastly, the effectiveness and performance of those algorithms in solving TSP across diverse scenarios and input parameters are compared.