APPLICATION OF ANT COLONY OPTIMIZATION TO

DEVELOP ENERGY EFFICIENT PROTOCOL IN

MOBILE AD-HOC NETWORKS

Sanjay K. Dhurandher

, Mohammad S. Obaidat

and Mayank Gupta

CAITFS, Division of Information Technology, Netaji Subhas Institute of Technology, University of Delhi, New Delhi, India

Department of Computer Science & Software Engineering, Monmouth University, New Jersey, U.S.A.

Division of Computer Engineering, Netaji Subhas Institute of Technology, University of Delhi, New Delhi, India

Keywords: Mobile ad hoc networks, Any colony schemes, Optimization, Energy-aware protocols.

Abstract: Foraging Behavior in Ant Swarms can be very helpful when applied to the protocols in Mobile Ad hoc

NETworks (MANETs). When the Ant Colony Optimization Scheme (ACO) is applied to a protocol, larger

number of paths are generated from the source to destination which helps in improving the packet delivery

ratio because an alternate back up path is always available in case a path gets broken due to the mobile

nodes. In this paper, we apply the ACO scheme on an already existing Energy efficient protocol Conditional

Max-Min Battery Capacity Routing (CMMBCR) (C.-K. Toh, 2001). The CMMBCR not only takes care of

the total transmission energy in the network but also the residual battery capacity of the nodes. Hence

applying ACO scheme on CMMBCR makes it more efficient in terms of energy, packet delivery ratio etc.

The efficiency of our proposed protocol A-CMMBCR is then established by comparing it with some of the

other existing Energy aware protocols such as Energy-Aware Routing protocol (EAAR) (Dhurandher et al.,

2009), Minimum Transmission Power Routing (MTPR) (Scott and Bambos, 1996) and CMMBCR. The

results are captured in the form of a graphical format.

1 INTRODUCTION

In the next generation of wireless communication

systems, there will be need of networks that can

establish themselves without any requirement of

preexisting infrastructure. Mobile Ad-Hoc Networks

(MANETS) basically refers to such type of networks.

As the name suggests Mobile implies that the

interconnecting nodes are not succumbed to be

remain at one place, rather they can move from one

place to the other. Ad-Hoc implies that the network

does not depend on any preexisting infrastructure

such as routers. Some of the main applications of

MANETS are dynamic communication for

emergency/rescue operations, disaster relief efforts

and military networks.

One of the most important performance

parameter in ad- hoc networks is minimizing the total

transmission energy in the path and extending the

battery life of the nodes. Conventional Routing

algorithms such as AODV (Perkins et al., 2001),

DSR (Johnson et al., 2001) and TORA (Park and

Corson, 2001) ignore the residual battery of the

participating nodes.

These protocols generally focus on finding the

shortest path available from source node to the

destination node. MTPR protocol tries to minimize

the total transmission power consumption of nodes

participating in an acquired route but it suffers from

the drawback that it does not consider the residual

battery of the nodes.

MMBCR (Singh et al., 1998) is another protocol

that finds the path which has longest battery life

amongst all other paths. CMMBCR is a combination

of MMBCR and MTPR. In this scheme, a parameter

gamma with some value assigned to it is used. Then

all paths from source node to destination node are

generated and the Minimum residual battery energy

(MBR) for each path is compared with the parameter

gamma. The paths which have MBR greater than

gamma are finally selected and MTPR scheme is

applied on this set of selected paths. In case no path

has MBR>gamma, the MMBCR scheme is followed.

Hence CMMBCR takes care of both

the residual

K. Dhurandher S., S. Obaidat M. and Gupta M..

APPLICATION OF ANT COLONY OPTIMIZATION TO DEVELOP ENERGY EFFICIENT PROTOCOL IN MOBILE AD-HOC NETWORKS.

DOI: 10.5220/0003677200120017

In Proceedings of the International Conference on Wireless Information Networks and Systems (WINSYS-2011), pages 12-17

ISBN: 978-989-8425-73-7

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

battery of the nodes as well as minimizing the total

transmission energy.

It has been seen from [8, 9, 10] that the Ant

Colony Optimization (ACO) scheme when applied to

ad-hoc networks greatly enhances the packet delivery

ratio. Some of the popular ACO based routing

schemes are AntNet (Dorigo et al., 1991), AntHocNet

(Di Caro et al., 2005) and ARA (Guenes et al., 2002).

The earlier ACO – based routing schemes such as

AntHocNet (Di Caro et al., 2005) and ARA (Guenes

et al., 2002), devised for ad-hoc networks, were not

targeted towards energy conservation.

In the protocol proposed in this paper, we have

applied this ACO scheme in CMMBCR. The

proposed protocol is inspired from the EAAR. In

EAAR, ACO scheme is applied on the already

existing MMBCR and thus significantly improving

the packet delivery ratio. In the proposed work we

have applied ACO scheme to develop a more

efficient energy aware protocol that not only takes

care of minimizing total energy consumed in the path

but also gives special attention to the residual battery

of nodes.

EAAR only talks about residual battery of nodes,

but doesn’t bother about the total transmission

energy. Our protocol is better than CMMBCR

because we have applied ACO scheme which ensures

that there is always a back up path available in case a

route breaks due to the mobile nodes. This greatly

enhances the packet delivery ratio. Moreover, it also

takes care of the fact that if a route gets overloaded

due to traffic, an alternate route is selected for

routing which ultimately takes care of the residual

battery of nodes.

2 PROPOSED SCHEME

Initially, when a Source node 'S' wishes to

communicate with a Destination node 'D' and it does

not have the routing information for ‘D’ available, it

broadcasts a route request packet (RREQ). Each

neighbour of ‘S’ thus receives the RREQ packet. At

each node this Request packet is used to find the

destination node and the corresponding node checks

whether there is an entry in its routing table for this

destination node. If an entry for the destination node

‘D’ is found the node sends a route reply packet back

to the source node along the same path from which it

received the RREQ. If it does not have any entry for

‘D’ available in its routing table, it further broadcasts

the RREQ packet. Furthermore, to apply the ACO

scheme, we need to calculate the pheromone for each

path. A-CMMBCR considers a combination of two

routing schemes, hence, we need to calculate two

pheromones – pheromone(mt) for MTPR and

pheromone(mm) for MMBCR.As the route request

packet traverses through the path it keeps on storing

the path so that the route request packet will have to

traverse back along the same path in the opposite

direction.

Meanwhile, all the route request packets received

get converted to route reply packets as soon as they

arrive to the destination and they travel back to the

source retracing the path. If this is not possible

because of the absence of the next hop due to node

movements, the route reply packet is discarded. At

the source node when RREP packet is received

corresponding values of pheromone(mt) and

pheromone(mm) are also received. Moreover the

MBR of that route is also received.

Each node has a routing table associated with it.

The routing table contains the addresses of

destination nodes along with the neighbor node to

which the source node should forward the packet in

order to make it reach the destination. Moreover it

contains the values of various pheromones associated

with a route.

If a source node 'S' wants to send data to a

destination node 'D' then following steps must take

place:

Step 1: The node S checks its routing table to find

whether a path to D exists or not. If a path exists, it

sends the data to the next Hop; else Step 2 is

performed.

Step 2: The node S broadcasts route request packet

(RREQ). Then Step 3 is performed.

Step 3: If any neighbor node’s routing table has a

path to D exists it replies back to node S through

Route Reply packet (RREP) else it broadcasts the

RREQ. Step3 is followed for each intermediate node

thus receiving the RREQ. If

no path for D is

available, the intermediate node relays the RREQ

packet.

Step 4: As the RREQ packet is broadcast in the

network, it can eventually reach the destination node

D. At the destination node, Route Reply packet

(RREP) is generated and reply is sent back to S.

RREP is passed to node S through the intermediate

nodes along the path from which RREQ was

received. Now as each node receives the RREP

packet, it updates its routing table and inserts an

entry for the destination node.

Calculations: Since we have to apply ACO we need

to know the Pheromone for each route generated and

our scheme requires calculation of TWO

pheromones: one for MTPR and the other for

APPLICATION OF ANT COLONY OPTIMIZATION TO DEVELOP ENERGY EFFICIENT PROTOCOL IN MOBILE

AD-HOC NETWORKS

MMBCR, which are calculated as follows:

Pheromone(mt])=1/(Total Transmission energy of

path * Number of Hops)

Pheromone(mm)= MBR/(Number of Hops)

where, MBR=Minimum battery of a node in the path.

Total transmission power is the sum of

transmission power to send data to next hop for each

node in the path.

We calculate MBR and Total Transmission

energy of path during the RREQ packet and

Pheromone(mt)and Pheromone(mm) during the

generation of RREP packet.

At the source node when RREP packet is received

corresponding values of pheromone(mt) and

pheromone(mm) are also received. Moreover the

MBR of that route is also received.

For all the routes obtained corresponding to a

particular destination we check:

if (MBR> γ)

{Select this route for MTPR}

else

{Select this route for MMBCR}.

For all those routes obtained for MTPR category the

route with highest Pheromone(mt) is selected for data

transmission. If no such route exists the Route with

highest Pheromone(mm) from MMBCR category is

selected for data transmission.

Assuming that the battery of any node has

maximum value of 100 units and applying ACO in

CMMBCR we get:

CMMBCR= ACO + MTPR if MBR>γ,

ACO + MMBCR otherwise

We can take value of gamma depending upon our

own choice.

Case 1: γ = 0

All routes will be selected for MTPR. Hence our

protocol performs similar to ACO+ MTPR

Case 2: γ =100

No route will be selected for MTPR and all routes

will be selected for MMBCR. Hence our algorithm

behaves as MMBCR+ACO.

Case 3: Taking any Random Value of γ between 0

and 100.

The proposed scheme will be followed

3 TEST CASES

Figure 1: An Illustrative example.

Note in Figure 1 the nodes are represented by circles

containing data in the form a:b , where a is node

address and b is the node battery level left. The data

on edges represents the power required to send data

between nodes forming the edge.

For convenience, the node battery level is taken

from 0 to 100 only.

From Figure 1 it is seen that there are 4 routes

from the source node ‘S’ to the destination node ‘D’.

These paths are listed below:

1. S -> 1 -> 2 -> 3 -> D

2. S -> 1 -> 2 -> 6 -> D

3. S -> 4 -> 5 -> 6 -> D

4. S -> 4 -> 5 -> 7 -> 8 -> D

For all these routes MBR, Pheromone(mm) and

Pheromone(mt) are calculated.

This data is shown for each of above 4 routes

below:

1. MBR=10, Pheromone(mm) =10/3,

Pheromone(mt) =

1/ (26*3).

2. MBR = 50, Pheromone(mm) =50/3,

Pheromone(mt) =

1/ (17*3).

3. MBR = 30, Pheromone(mm) =30/3,

Pheromone(mt) =

1/ (39*3).

4. MBR = 30, Pheromone(mm)=30/4,

WINSYS 2011 - International Conference on Wireless Information Networks and Systems

Pheromone(mt) =

1/ (47*4).

Now depending on the value of γ different routes can

be selected for data transmission using MTPR or

MMBCR.

In this test case:

If γ<= 49, route 2 will be selected for transmission

using Pheromone(mt).

Else route 2 will be selected for transmission using

Pheromone(mm).

The other routes can also be used for data

transmission by comparing their MBR with the value

of γ and deciding whether to use MMBCR or MTPR.

4 SIMULATION ANALYSIS AND

RESULTS

In this section, we report the results generated by

conducting the simulation experiments and

comparing our protocol with some standard and

selected benchmark protocols. Simulation was done

using Glomosim tool.

The following parameters were considered for the

simulations performed:

1) Simulation time: 500 seconds

2) Terrain dimensions: (2000,2000) meters square

3) Number of Nodes:30

4) Mac Protocol: 802.11

5) Initial energy of Nodes: All Nodes were initiated

with equal energy.

The traffic considered n this work is the Constant Bit

Rate (CBR) traffic with the following scenarios:

1) CBR 17 100 1536 1S 0S 250S

2) CBR 12 19 100 1536 1S 250S 400S

3) CBR 14 27 100 1536 1S 400S 500S

The benchmark protocols used to compare with our

protocol are CMMBCR, EAAR, and MTPR.

Since our protocol is an improvement over

CMMBCR hence we decided to take this protocol in

our consideration. EAAR is an improvement our

MMBCR in that it implements the ACO scheme.

MTPR tries to optimize the energy used in the

network. Hence the choice of benchmark protocols is

justified.

We conducted the simulation experiments under

the following six situations:

1) Data size =100 times Control Packet Size;

Mobility :NONE

2) Data size =125 times Control Packet Size;

Mobility :NONE

3) Data size =150 times Control Packet Size.

Mobility :NONE

4) Data size =100 times Control Packet Size;

Mobility speed: 10m/s, random way point model.

5) Data size =125 times Control Packet Size;

Mobility speed: 10m/s, random way point model.

6) Data size =150 times Control Packet Size;

Mobility speed: 10m/s, random way point model.

Parameters that we considered for comparison with

other protocols are: (1) Total Energy consumed, (2)

Number of dead nodes, (3) Number of packets

delivered, (4) Energy per packet delivered, and (4)

Number of packets

dropped.

Figure 2: Energy consumed per packet.

Figure 2 shows that the energy consumed per

packet in the network is the least for A-CMMBCR.

A-CMMBCR performs better than CMMBCR

because the ACO scheme generates multiple paths

from one node to the other. When the traffic on one

path increases its pheromone decreases. This is done

so that the packets that would be transmitted later on

would go through some different path rather than

overloading this path. This helps in increasing the

number of packets delivered and hence lesser energy

is consumed per packet.

A-CMMBCR performs better than EAAR here

because the energy consumed in the network is lesser

as the path having the highest pheromone consumes

lower energy than the normal path selected by ACO

scheme in EAAR.

APPLICATION OF ANT COLONY OPTIMIZATION TO DEVELOP ENERGY EFFICIENT PROTOCOL IN MOBILE

AD-HOC NETWORKS



Figure 3: Number of Packets Delivered.

In the first scenario mobility is set to zero. Hence,

the nodes do not move, which implies that once a

path between two nodes has been established it

would remain intact. Number of packets delivered

using ACO scheme are higher than the normal on

demand scheme as shown in figure 3.



Figure 4: Number of Packets Dropped.

Figure 4 shows the number of packets dropped by

each protocol for various scenarios. It can be seen

that the protocol based on ACO scheme has lesser

number of packets dropped; reason being that there is

always an alternative path available in case the

current path gets broken or is overloaded with traffic.

This guarantees that most of the packets would reach

the destination more often than not.

Figure 5 shows that CMMBCR and A-CMMBCR

has lesser number of dead nodes than other protocols.

This is due to the use of a combination of MTPR and

EAAR, which ensures that the least energy would be

used in the network along with taking care of weak

nodes. Hence the probability of choosing a path that

has weak node is very low.

From the above results, it has been seen that A-

CMMBCR performs better than other three protocols

because it is based on ACO scheme. Moreover, ACO

scheme does guarantee the availability of multiple

paths for data transfer, which ensures a higher packet

delivery ratio.



Figure 5: Number of Dead Nodes.

Figure 6: Total Energy Consumed in the Network.

5 CONCLUSIONS AND FUTURE

WORK

This paper has presented an ACO based A-

CMMBCR energy efficient routing technique to

conserve energy in the process of routing of data

from one node to another. Furthermore, from the

graphical results it can be concluded that the

proposed A-CMMBCR performs better than the

CMMBCR protocol as the overall energy consumed

in the network is reduced and also the number of

packets dropped is decreased due to the ACO scheme

applied on the CMMBCR. The results also point

towards the better performance of the A-CMMBCR

in terms of energy over the other energy efficient

protocols, namely, EAAR and the MTPR of which

EAAR is the most recently designed/proposed and is

also based on the ACO technique.

One of the limitations of the proposed A-

CMMBCR is that it lacks the fault tolerance aspect.

So, our next work would be to ensure that the

network selects only those paths in which the nodes

are not prone to any fault and if there exists no such

WINSYS 2011 - International Conference on Wireless Information Networks and Systems

path then select those paths which are least prone to

faults. The new protocol would make sure that the

selected path has highest fault tolerance amongst the

other paths.

REFERENCES

C.-K. Toh, “Maximum Battery Life Routing to Support

Ubiquitous Mobile Computing in Wireless Ad Hoc

Networks”, IEEE Communications Magazine, Vol. 39,

No. 6, June 2001, pp. 138-147.

Sanjay K. Dhurandher , Sudip Mishra and Mohammad S.

Obaidat , “ An Energy-Aware Routing Protocol for Ad-

Hoc Networks Based on Foraging Behavior in Ant

Swarms,” IEEE ICC 2009, pp. 1-5.

K. Scott and N. Bambos, “Routing and Channel

Assignment for Low Power Transmission in PCS”,

Proc. Intl. Conf. Universal Personal Communications

(ICUPC’96), Cambridge, MA, 1996, pp. 498-502.

C. E. Perkins, E. M. Belding-Royer, and S. Das, “Ad Hoc

On-demand Distance Vector (AODV) Routing”, IETF

Internet Draft, 2001.

B. Johnson, D. A. Maltz, Y.-C. Hu and J. G. Jetcheva,

“The Dynamic Source Routing for Mobile

AdHocWirelessNetworks”,http://www.ietf.org/internet

-drafts/draft-ietfmanet-dsr-06.txt,IETF Internet Draft,

Nov. 2001.

V. Park and S. Corson, “Temporally-Ordered Routing

Algorithm (TORA) Version 1”, IETF Internet Draft,

July 2001.

S. Singh, M. Woo, and C. S. Raghavendra, “Power-Aware

Routing in Mobile Ad Hoc Networks”, Proc. 4th

Annual ACM/IEEE Intl. Conf. Mobile Computing and

Networking, 1998, Dallas, TX, pp. 181-190.

M. Dorigo, V. Maniezzo and A. Colorni, The Ant System:

An Autocatalytic Optimizing Process, TR91-016,

Politecnico di Milano, 1991.

G. Di Caro, F. Ducatelle and L. M. Gambardella,

“AntHocNet: An Adaptive Nature-Inspired Algorithm

for Routing in Mobile Ad Hoc Networks”,

Telecommunications (ETT), Vol. 16, No. 2, 2005.

M. Guenes, U. Sorges and I. Bouazizi, "ARA: The Ant-

Colony Based Routing Algorithm for MANETs", Proc.

of ICPPW'02, 2002, pp. 79- 85.

APPLICATION OF ANT COLONY OPTIMIZATION TO DEVELOP ENERGY EFFICIENT PROTOCOL IN MOBILE

AD-HOC NETWORKS