Modified Gorilla Troops Optimizer with Elite Opposition-Based Learning and Tangent Flight Operator to Solve Traveling Salesman Problem

Abstract

Artificial Gorilla Troops Optimization (GTO) is inspired by the foraging and social behavior of gorillas but tends to get trapped in local optima and experiences premature convergence when applied to the Traveling Salesman Problem (TSP). GTO belongs to a group of optimization methods known as metaheuristics, which are high-level algorithms designed to find near-optimal solutions for complex optimization problems by combining global exploration and local exploitation strategies. Metaheuristics are particularly effective for solving problems where traditional exact methods are inefficient or impractical due to large search spaces. To overcome these limitations, Modified Gorilla Troops Optimization (mGTO) was developed by incorporating Elite Opposition-Based Learning (EOBL) and the Tangent Flight Operator (TFO). The EOBL enhances the diversity of the population which uses elite individuals to generate opposite solutions, leveraging their superior information to refine particle positions and effectively guide the population toward high-potential regions in the search space. The TFO adjusts each solution's position toward the best solution using tangent-based random steps and impact force to enhance local search and avoid premature convergence. In this study, mGTO was evaluated against several algorithms including well-known classic metaheuristics, the latest metaheuristics, and variants of GTO. The mGTO was executed on several instance datasets of TSP and tested using Friedman's non-parametric & Holm's test with outperforming results compared to 11 algorithms. These findings demonstrate the effectiveness of mGTO in providing improved routing solutions and highlight its potential for enhancing sustainable logistics by minimizing travel distances.