International Journal of Industrial Engineering Computations 7 (2016) 245–256
Contents lists available at GrowingScience
International Journal of Industrial Engineering Computations
homepage: www.GrowingScience.com/ijiec
A population-based algorithm for the multi travelling salesman problem
Rubén Iván Bolaños
a
, Eliana M. Toro O
b
and Mauricio Granada E
a*
a
Department of Electrical Engineering, Universidad Tecnológica de Pereira, Colombia
b
Department of Industrial Engineering, Universidad Tecnológica de Pereira, Colombia
C H R O N I C L E A B S T R A C T
Article history:
Received July 22 2015
Received in Revised Format
Septmber17 2015
Accepted October 14 2015
Available online
October 15 2015
This paper presents the implementation of an efficient modified genetic algorithm for solving
the multi-traveling salesman problem (mTSP). The main characteristics of the method are the
construction of an initial population of high quality and the implementation of several local
search operators which are important in the efficient and effective exploration of promising
regions of the solution space. Due to the combinatorial complexity of mTSP, the proposed
methodology is especially applicable for real-world problems. The proposed algorithm was
tested on a set of six benchmark instances, which have from 76 and 1002 cities to be visited. In
all cases, the best known solution was improved. The results are also compared with other
existing solutions procedure in the literature.
© 2016 Growing Science Ltd. All rights reserved
Keywords:
Combinatorial optimization
Multi-traveling salesman problem
Population-based algorithm
Local search operators
1. Introduction
The mTSP problem can be viewed from the perspective of two well-known problems: i) as a
generalization of the Travelling Salesman Problem (TSP), where a set of routes is assigned to m salesmen
who all start from and return to a home city, and ii) as a special case of the vehicle routing problem
(VRP), in which customers are considered unitary demands and every travelling salesman only visits a
predetermined number of cities. Thus, the mTSP can also be utilized for solving several types of VRPs
and all formulations and solution approaches for the VRP are valid for the mTSP.
Although the VRP and TSP have been widely discussed in the literature, the research on the mTSP is
limited. Moreover, few papers in the literature address the mTSP through efficient population-based
algorithms. The main motivation for formulating a population methodology lies in the ease of integration
with multi-objective strategies, which allow introducing practical aspects, such as profit, fuel
consumption and environmental impact, among others. Therefore, it becomes relevant to develop and
implement an effective and robust optimization methodology based on population.
* Corresponding author.
E-mail: magra@utp.edu.co (M. Granada E)
© 2016 Growing Science Ltd. All rights reserved.
doi: 10.5267/j.ijiec.2015.10.005