Cellular Competitive Decision Algorithm for Traveling Salesman Problem
Xiaohua Xiong
College of Computer and Information
Shanghai Second Polytechnic University
Shanghai, China
e-mail: xhxiong@sspu.edu.cn
Aibing Ning
School of Management
University of Shanghai for Science and Technology
Shanghai, China
e-mail: nab@usst.edu.cn
Abstract— Among all combinatorial optimization problems,
traveling salesman problem (TSP) is one of the widely studied
problems. Though the optimization version of this problem is
NP-hard, practical solution techniques don’t require
optimality. And many different heuristics and approximation
algorithms are used to solve this problem. Cellular competitive
decision algorithm (CCDA) is a new heuristic for solving hard
problem, especially those NP-hard problems. One of the
biggest characteristic of CCDA is easy to integrate the
mathematical feature with the problem itself together. Firstly,
mathematical properties of traveling salesman problem are
analyzed to calculate the lower bound of the problem. Then a
CCDA for traveling salesman problem is presented. In order to
test its validity, we carry out extensive computational
experiments. From the results we know the presented
algorithm has a satisfactory behavior both in running time and
the quality of the solution found.
Keywords- traveling salesman problem; cellular automaton;
lower bound;upper bound; algorithm.
I. INTRODUCTION
Among all the combinatorial optimization problems,
traveling salesman problem (TSP) is one of the most widely
studied problems. There is a list of cities and the distances
between each pair of cities. Now there is a salesman who
wants to go to all cities to sell his goods. He needs to find a
shortest route to visit each city exactly once and return to the
original city [1-2]. It was first founded by Edmonds, Cook
and Karp in 1930. It is an NP-complete problem, which
means there is no exact algorithm to solve large-scale
instances of TSP [2]. For the running time of exact algorithm
will increase exponentially and fail to give a feasible solution
with the number of cities increase. Due to its theoretical and
practical values, TSP is one of the most intensively studied
problems in the fields of combinatorial optimization,
operation research and computer science in order to resolve
it within a reasonable computation time. TSP has many
applications in its purest formulation, such as city traffic
planning, logistics transportation, communication network
and etc. And it also appears as a sub-problem in many areas,
such as DNA sequencing [2].
Let’s describe TSP in mathematical language. There is a
graph G=(V, E). V={1,2,…,n} is the set of nodes. And E is
the set of edges. For i=1,…, n, d
ij
is the distance from node i
to node j. Let
1 edge ( , ) is in optimal route
0otherwise
ij
ij
x
=
®
¯
Then the mathematical model of TSP can be formulated
as follows:
{}
11
1
1
(1)
1, (2)
1, (3)
s.t.
1, (4)
0,1
nn
ij ij
ij
n
ij
j
n
ij
i
ij
iSjS
ij
MinZ d x
xiV
xjV
xS SV
x
==
=
=
∈∈
=
¦¦
=∈
¦
=∈
¦
≤−∀⊂
¦¦
∈ (5)
°
°
°
°
®
°
°
°
°
¯
Constraint 2 and constraint 3 requires that each node can
be arrived from exactly one other node and can depart to
exactly one other node. The fourth constraint makes sure that
there is no any sub-route solution.
Enumeration method is the most direct solution for TSP.
But the running time of it is O(n!), it is impossible even the
city’s number is 20. There are some other exact algorithms,
such as cutting plane method [3], branch-and-bound
algorithm [4]. Exact approaches can guarantee the optimality
of the solution. But they couldn’t get a feasible solution
because of the running time is enormous big especially in
some engineering solutions. Though heuristics and
approximation algorithms couldn’t guarantee optimality,
they can used to generate a feasible, suboptimal solution
when the case optimality is not so necessary [5]. Various
heuristics and approximation algorithms, which demand
heuristic strategies in tour construction and tour
improvement, have been devised [6-10]. Generally, the tour
construction strategy is used in the nearest neighbor (NN)
algorithm and the variation of NN algorithm. The tour
improvement method is widely used in many heuristics, such
as tabu search algorithm, ant algorithm, genetic algorithm,
particle swarm algorithm and the cross entropy method. In
this paper, we present a cellular competitive decision
algorithm (CCDA) for TSP. Most of heuristics have a
common characteristic is not considering the deviation
degree between the solution and the optimal solution. For
one way, it is hard to get it. On the other way, it needs
profound mathematical knowledge and skills. We will
research on the mathematical properties of TSP and analysis
the lower bound of TSP. And it can be used to judge the
2014 Second International Conference on Enterprise Systems
978-1-4799-5554-1/14 $31.00 © 2014 IEEE
DOI 10.1109/ES.2014.54
219
2014 Second International Conference on Enterprise Systems
978-1-4799-5554-1/14 $31.00 © 2014 IEEE
DOI 10.1109/ES.2014.54
221
2014 Second International Conference on Enterprise Systems
978-1-4799-5554-1/14 $31.00 © 2014 IEEE
DOI 10.1109/ES.2014.54
221