Redirigiendo al acceso original de articulo en 21 segundos...
Inicio  /  Algorithms  /  Vol: 14 Par: 9 (2021)  /  Artículo
ARTÍCULO
TITULO

A New Constructive Heuristic Driven by Machine Learning for the Traveling Salesman Problem

Umberto Junior Mele    
Luca Maria Gambardella and Roberto Montemanni    

Resumen

Recent systems applying Machine Learning (ML) to solve the Traveling Salesman Problem (TSP) exhibit issues when they try to scale up to real case scenarios with several hundred vertices. The use of Candidate Lists (CLs) has been brought up to cope with the issues. A CL is defined as a subset of all the edges linked to a given vertex such that it contains mainly edges that are believed to be found in the optimal tour. The initialization procedure that identifies a CL for each vertex in the TSP aids the solver by restricting the search space during solution creation. It results in a reduction of the computational burden as well, which is highly recommended when solving large TSPs. So far, ML was engaged to create CLs and values on the elements of these CLs by expressing ML preferences at solution insertion. Although promising, these systems do not restrict what the ML learns and does to create solutions, bringing with them some generalization issues. Therefore, motivated by exploratory and statistical studies of the CL behavior in multiple TSP solutions, in this work, we rethink the usage of ML by purposely employing this system just on a task that avoids well-known ML weaknesses, such as training in presence of frequent outliers and the detection of under-represented events. The task is to confirm inclusion in a solution just for edges that are most likely optimal. The CLs of the edge considered for inclusion are employed as an input of the neural network, and the ML is in charge of distinguishing when such edge is in the optimal solution from when it is not. The proposed approach enables a reasonable generalization and unveils an efficient balance between ML and optimization techniques. Our ML-Constructive heuristic is trained on small instances. Then, it is able to produce solutions?without losing quality?for large problems as well. We compare our method against classic constructive heuristics, showing that the new approach performs well for TSPLIB instances up to 1748 cities. Although ML-Constructive exhibits an expensive constant computation time due to training, we proved that the computational complexity in the worst-case scenario?for the solution construction after training?is O(n2logn2)" role="presentation">??(??2log??2)O(n2logn2) O ( n 2 log n 2 ) , n being the number of vertices in the TSP instance.

 Artículos similares

       
 
Daniela Ambrosino, Carmine Cerrone and Anna Sciomachen    
This paper presents a new heuristic algorithm tailored to solve large instances of an NP-hard variant of the shortest path problem, denoted the cost-balanced path problem, recently proposed in the literature. The problem consists in finding the origin?de... ver más
Revista: Algorithms

 
Carlos Bautista-Capetillo, Georgia Aralú González Pérez, Hiram Badillo Almaraz and Aldo López Valle    
Civilizations have been able to bloom because of the way they have been historically associated with water resources, especially in seeking strategies to ensure a supply to diverse sectors that require them. Thus, challenges in satisfying water demand ar... ver más
Revista: Water

 
Kamila Klimek, Karol Postawa, Magdalena Kaplan and Marek Kulazynski    
Great interest in viticulture in temperate climates results from the introduction of new interspecies hybrids of grapevines which are quite popular due to their high resistance to fungal diseases and lower temperature. However, the impact of rootstocks, ... ver más
Revista: Applied Sciences

 
Ziming Zhang, Xinping Wang, Chang Su and Linhui Sun    
Quality improvement is crucial for manufacturing, and existing research has paid less attention to the influence of regulatory factors and irrational factors of decision makers. Considering the impact of the reward and punishment strategy of the shared p... ver más
Revista: Applied Sciences

 
Abdullah Khan, Imran Shah, Shahid Aziz, Muhammad Waqas, Uzair Khaleeq uz Zaman and Dong-Won Jung    
The bullet head plays a principal role in the modern enlargement of an efficient bullet. A bullet?s main design parameters depend upon the lift and drag forces acting on the head. The factors in a bullet?s shape design that affect bullets? lift and drag ... ver más
Revista: Aerospace