Page 1

A Meta-heuristic Method for Cluster-based Channel

Assignment in a Mobile Ad Hoc Network

Mahboobeh Parsapoor and Urban Bilstrup

School of Information Science, Computer and Electrical Engineering (IDE), Halmstad University.

mahpar11@student.hh.se, urban.bilstrup@hh.se.

Abstract—In this paper, a centralized algorithm on the basis

of ant colony optimization (ACO), a well-known bio-inspired

model is suggested as a channel assignment scheme. The

suggested method is applied for a cluster-based MANET and

aims at minimizing the number of used channels while

satisfying the co-channel interference constraints. The

suggested algorithm is examined for two scenarios and the

obtained results are compared with a GA-based (genetic

algorithm) scheme.

Index Terms—Ant Colony Optimization, Channel

Assignment Problem in MANET, Co-channel Interference,

Spectral Efficiency.

I. Introduction

Next generation military wireless communication systems at

the tactical edge will be based on mobile ad hoc networks

(MANETs). In the case of large military endeavors

extremely high channels utilization is necessary and this can

only be achieved if efficient spatial channels reuse is

possible. Finding an efficient channel reuse has been

classified as the channel assignment problem [1] that was

early defined as a frequency assignment problem in cellular

communication systems [1]-[4]. In frequency assignment

problem, a feasible scheme should maximize spatial channels

reused while satisfying interference constraints (e.g. co-

channels interference constraint). It should also address

several issues, such as: stability, throughput, connectivity

and fault tolerance [1]-[4].

For different types of wireless network architectures,

numerous heuristic methods have been proposed for

solving channel assignment problem that has been

identified as an NP-hard problem [5]-[8].

This paper suggests the use of ant colony optimization

(ACO) that is a meta-heuristics method and has been

inspired by ants’ behavior. The rest of the paper is

organized as follows: Section II reviews the ant colony

optimization method. Section III describes the obtained

results of applying an ACO-based channel assignment

scheme for three MANETs. Section IV presents some

significant notes and explains about the future works.

II. Ant Colony Optimization: A Meta-

heuristic Methods

Recent researches have shown that the meta-heuristic

methods are more efficient than the classical optimization

methods (e.g., branch and bound) to solve NP-hard

problems. Ant colony optimization is a well-known meta-

heuristic method that has been shown excellent

performance on solving NP-hard problem. Ant colony

optimization has been inspired from the behavior of real

ants to construct the shortest path from the nest to the

source of food. To solve an optimization problem using an

ACO-based algorithm, the problem is represented by a

graph. A best solution is a sequence of the graph nodes

with minimum cost function. This algorithm stars with a

population of ants that are randomly placed on the graph

nodes. To traverse the graph, each ant chooses the next

node using a ‘probabilistic transition rule’ [9],[10] that

consists of two components: a heuristic function and

pheromone intensity. The heuristic function is a problem

independent function and shows the desirability of the next

node, while pheromone intensity represents the desirability

of the next path from the perspective of the best ant (the ant

that finds the path between the source and destination with

a minimum cost function). The cost function is assigned to

each complete path between the source and destination

node and indicates how profitable the path is. In channel

assignment problem for a clustered MANET, the graph of

problem

) ' , '('

EVG

consists of nodes, '

the cluster heads and links, ' E , that represent that the two

clusters are neighbors, i.e., they have a common node

(gateway node). For finding a solution, ants should traverse

this graph and choose different channels for neighbor

cluster heads.

V , that represent

III. Simulation Results

We use the ACO-based method for solving the channel

assignment problem in two simulated scenarios using

MATLAB. The results are compared with the results of a

GA-based method that is proposed in [11]. We have taken

the following assumptions: 1) A centralized base station

senses the available channels and decides for a channel

assignment scheme. 2) In the arrival time of demands for

channels, an assignment scheme is provided on the basis of

the available channels. 3) The topology of network and the

transmission power of nodes are assumed to be static

Page 2

during the execution of the channel assignment algorithm.

4) All of nodes of network are assumed to have same

transmission power. In the first scenario, a MANETs with

N nodes (i.e., N=50 and 100) and different transmission

ranges (i.e., TR=100, 200 and 300 meters) are simulated.

Lowest ID (LID) clustering is applied to form the clustered

network structure. The nodes are placed in a 1000 x 1000

meter square and the position of each individual node has

two coordinates, x and y, that are drawn from a uniform

distribution. The number of used channels to the MANETs

with different nodes is depicted in Figure 1. Obviously, the

number of required channels is dependent upon the number

of nodes and the topology of the network. However, for the

same topology, the number of assigned channels by an

ACO-based method is smaller than a GA-based method.

(See Figure 1 the black and grey bars). It can also be

observed using ACO-based method the number of assigned

channels has no significant change when the size of the

network (the number of clusters) increases. Thus, the

methods might be scalable for a large sized MANET.

100200 300

0

5

10

15

20

25

30

Transmission range

Average Numbers

No.Assigned channel ACO based Scheme

No. Assigned channel GA based Scheme

No.Clusters

(a)

100200300

0

5

10

15

20

25

30

35

Transmission range

(b)

Figure 1. The average number of assigned channels and clusters. (a).

for a network with 50 nodes, (b). for a network with 100 nodes.

For the second scenario, a clustered MANET with 20

clusters is examined. The simulated MANET has 100 nodes

and the transmission range and interference range are

assumed to be 100 and 250 meters, respectively. The

convergence characteristics of the ACO and GA algorithms

are depicted in Figure 2. It can be observed that ACO-based

method converges after approx. 50 iterations, while the GA-

based algorithm converges after 60 iterations. However, the

average and minimum values of objective functions using

GA differs to a large extent and it converges much slower to

reach a minimum of the objective. .

IV. Conclusion

In this paper, a cluster-based channel assignment scheme

on the basis of ant colony optimization method is proposed

and examined by two scenarios. We also compare its results

with the obtained results from a GA-based scheme. The

obtained results show that the proposed ACO has the

capability to approximate the solution to minimize the

average level of assigned channels. The results also

indicate that the performance of an ACO-based channel

allocation scheme is not dependent on the size of MANET,

e.g., the number of clusters in MANET. Thus, it provides a

stable and scalable scheme. In future, we will develop a

distributed clustered-based scheme on the basis of ACO

and replace the present lowest ID clustering algorithm with

an ACO-based clustering method to effectively address the

channel assignment problem.

0

1020 3040506070

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Iteration

Minimum of Objective Function ACO

Average of Objective Function ACO

Minimum of Objective Function GA

Average of Objective Function GA

Figure 2. The minimum and mean values of cost function versus the number of

iterations for GGA and ACO.

References

[1] G. K. Audhya, K. Sinha, S. C. Ghosh, and B. P. Sinha, “A survey

on the channel assignment problem in wireless networks,”

Wirel. Commun. Mob. Comput., vol. 11, pp.583-609, 2010.

[2] G. K. Audhya, K. Sinha, S. C. Ghosh, and B. P. Sinha, “A survey

on the channel assignment problem in wireless networks,”

Wirel. Commun. Mob. Comput., vol. 11, pp.583-609, 2010.

[3] T. C. Hou, C. M. Wu and M. C. Chan. “Dynamic channel

assignment in clustered multi hop CDMA/TDMA ad hoc

networks,” in Proc. PIMRC 2004. vol.1, pp. 145- 149, 2004.

[4] S. H. Wong. “Channel Allocation for Broadband Fixed Wireless

Access Networks” PhD dissertation, University of Cambridge,

Department of Engineering, UK, 2003

[5] V. Maniezzo, A. Carbonaro, “An ANTS heuristic for the

frequency assignment problem,” Future. Gener. Comp. Sy, vol.

16, pp. 927-935, 2000.

[6] T.C. Hou and T.J. Tsai, “On the Cluster Based Dynamic

Channel Assignment for Multi hop Ad Hoc Networks,” J.

Commun. Netw., vol. 4, no. 1, pp. 40–47, 2002.

[7] S. H. Wong and I. Wassell, “Dynamic channel allocation using a

genetic algorithm for a T DD broadband fixed wireless

access network,” in Proc. IASTED, , pp. 521–526, 2002.

[8] P. M. Papazoglou, D. A. Karras, and R. C. Papademetriou, “On

the Implementation of Ant Colony Optimization Scheme for

Improved Channel Allocation in Wireless Communications,” in

Proc. IS '08, vol.1, no., pp.6-41-6-50, 2008.

[9] M. H. Jin, H. K. Wu, J. T. Horng and C. H. T, “An

evolutionary approach to fixed channel assignment problems

with limited bandwidth constraint,” in Proc. ICC2001, vol.7,

pp.2100-2104, 2001.

[10] T. S. C. Felix, M. K. Tiwari, Swarm Intelligence: Focus on Ant

and Particle Swarm Optimization, I-Tech Education and

Publishing, 2007.

[11] A. P. Engelbrecht, Fundamentals of Computational Swarm

Intelligence, John Wiley & Sons, 2005.

[12] M. Parsapoor, U. Bilstrup, "Using the grouping genetic

algorithm (GGA) for channel assignment in a cluster-based

mobile ad hoc network," in Proc. SNCNW 2012, 2012.