Optimization of the traveling salesman problem through a genetic algorithm guided by self-organizing maps

Abstract

Optimization is a fundamental process for achieving objectives, exemplified by the Traveling Salesman Problem (TSP), which minimizes resource consumption. This study focuses on optimizing travel routes to reduce fuel costs and maximize tourist visits through efficient time management. We employ a genetic algorithm (GA), a powerful optimization technique, to determine routes. A key challenge in GAs is premature convergence, which can prevent optimal solutions. To address this, we introduce a novel approach integrating Self-Organizing Maps with GAs, specifically through a new selection operator designed to enhance GA performance. An application was developed using real-world coordinates of Pakistani cities. Simulation results demonstrate that our proposed method significantly outperforms existing selection operators in terms of final best distance, average best distance, standard deviation, and computational time. Field validation further confirmed an 18.3% distance reduction and 4.53 million in annual savings in a logistics case study.