Appl. Math. Inf. Sci. 7, No. 2, 777-784 (2013) 777
Applied Mathematics & Information Sciences
An International Journal
c
2013 NSP
Natural Sciences Publishing Cor.
A Novel Discrete Cuckoo Search Algorithm for Spherical
Traveling Salesman Problem
Xinxin Ouyang
1
, Yongquan Zhou
1,2
, Qifang Luo
1
and Huan Chen
1
1
College of Information Science and Engineering, Guangxi University for Nationalities, Nanning Guangxi 530006 P.R. China
2
Guangxi Key Laboratory of Hybrid Computation and IC Design Analysis, Nanning 530006, P.R. China
Received: 20 Sep. 2012; Revised 26 Oct. 2012; Accepted 28 Nov. 2012
Published online: 1 Mar. 2013
Abstract: In this paper, we propose a novel discrete cuckoo search algorithm (DCS) for solving spherical Traveling Salesman Problem
(TSP) where all points are on the surface of a sphere. The algorithm is based on the L
´
evy flight behaviour and brood parasitic behaviour.
The proposed algorithm applies study operator, the ”A” operator, and 3-opt operator to solutions in the bulletin board to speed up the
convergence. Optimization results obtained for HA30 (an instance from TSPLIB) and different size problems are solved. Compared
with GA, DCS is better and faster.
Keywords: discrete cuckoo search algorithm (DCS), spherical geometry, spherical TSP
1. Introduction
The Traveling Salesman Problem (TSP) is classical and
most widely studied problem in the field of Combinato-
rial Optimization and attracts computer scientists, mathe-
maticians, and others [1]. The idea of problem is to find
shortest route of salesman starting from a given city, vis-
iting n cities only once and finally arriving at origin city.
Euclidean TSP is a NP-hard problem as the search space
is huge viz. n!, so that it is calculating infeasible to ob-
tain optimal solution to the problem. [2] used local search
operators best part collect, C2Opt, smallest square evolu-
tion for TSP, the results perform good. [3] Combined the
GGA and LGA as an operator to solve TSP. [4] proposed
a greedy heuristic algorithm with regret and the cheapest
insertion algorithm for TSP.
In metric TSP the nodes lie in a metric space (i.e., the
distances satisfy the triangle inequality), whereas in Eu-
clidean TSP the nodes lie in R
2
(or more generally, in R
4
for each d). Much of work on TSP lie in R
2
has been
done with heuristic algorithms to produce an optimal or
close to optimal solution. The commonly used heuristic
approaches are greedy algorithms; 2-opt, 3-opt[1]; simu-
lated annealing(SA); ant colony algorithm(ACA)[5]; ge-
netic algorithm (GA); particle swarm optimization (PSO)[6];
Artificial neural network (ANN) [7]. While the work on
multi dimensional TSP is seediness [8] solved 3D-TSP
for the multi-dimensional city location. [9] solved TSP
with genetic algorithm on a sphere. [10] has proved TSP
S
(TSP on the curve) is NP hard and can be polynomial time
approximated. [11] has described the distribution of dis-
tances between random points on a sphere. [12] gave an
upper bound and a lower bound of the optimal value to the
random spherical TSP.
In this paper, we propose a novel discrete cuckoo search
algorithm (DCS) for solving the Euclidean TSP, where all
points are on the surface of a sphere. The cuckoo search
(CS) algorithm proposed by Yang and Deb (2009, 2010) is
based on L
´
evy flight behavior and brood parasitic behavior
[13–15]. CS performed excellent on function optimization,
engineering design, training neural network and other con-
tinuous target optimization problems [16–18]. Now CS is
proposed to solve knapsack problem and nurse scheduling
problem [19,20].
The proposed algorithm does not save the initial routes
in the bulletin board, until the segment between each city
and its around cities partially reversed. In every genera-
tion, every cuckoo searches a new nest and abandons the
old one to decrease the TSP route. In order to accelerate the
convergence, the proposed algorithm applies study opera-
∗
Corresponding author: e-mail: yongquanzhou@126.com
c
2013 NSP
Natural Sciences Publishing Cor.