Machine Learning Approaches for the Traveling Salesman Problem: A Survey

Abstract

Machine Learning techniques have been applied in many contexts with great success. In this survey, we focus on their applications in the Combinatorial Optimization (CO) domain, and in particular to the Traveling Salesman Problem (TSP). We propose an intuitive and simple mind map helpful to navigate through the wide existing literature, indicating the approaches we consider most promising. Different ML techniques introduced to solve the TSP are discussed and reviewed; and their differences and limitations are delved. Open problems for future research in this area are finally highlighted.