EAI Endorsed Transactions

on __________________________ Research Article

Metaheuristics to aid Energy-Efficient Path Selection in

Route Aggregated Mobile Ad hoc Networks
Deepa Mehta1,1, Sherin Zafar2 and Indu Kashyap3

1
Assistant Professor, DEEE, GD Goenka University, Sohna(deepa.mehta@gdgoenka.ac.in)
2
Assistant Professor, SEST, Jamia Hamdard, New Delhi(sherin)
3
Professor, FET, MRIIRS, Faridabad

Abstract

The field of Information Technology is buzzing with the new paradigm “Internet of Things (IoT)”. Being an element
of the Future Internet, IoT performs integration of technologies and communication solutions, the envisioned
connectivity of heterogeneous devices procreates routing challenges in IOT and it becomes important to specify how
the current standard communication protocols can provide support for this vision realization. This opens a new
research horizon for MANET protocols. New applications associated with the Internet of Things (IoT) often involve
self-organization and collaboration among many autonomous wireless devices for which traditional infrastructure-
based wireless networks are not suitable. Mobile Ad hoc Networks (MANETs), with their infrastructure-less,
spontaneous and arbitrary multi-hop features are considered a popular approach for such scenarios, although they
also face certain inherent challenges. Energy efficient routing is one such area of interest, which require non-
conventional paradigm for design and development of protocols. Swarm Intelligence (SI) based metaheuristic owing
to their self-organising nature, prove to be an efficient solution for routing problems, to deliver events in an energy
constraint MANET environment. In this paper, we present, a SI based, energy-efficient hybrid routing protocol for
MANETs. The proposed protocol consists of three phases: 1) Cluster Design phase, 2) ACO aided cluster head
selection phase and 3) Data Transmission using route aggregation phase. Integration of three phases for node
clustering, data routing and transmission, is the key aspect of our proposed protocol, which ultimately contributes to
its robustness. Evaluation of simulation results show that SI perform better in terms of packet delivery, energy
consumption and throughput with increased network life compared to the conventional hybrid routing protocols. One
of the direct influences of routing protocol is the performance of route aggregation in MANET. This paper presents a
novel routing protocol that performs a cost effective, reliable and robust routing mechanism in MANET with its
highly compatibility with in-network route aggregation mechanism.

Keywords: Route Aggregation, Ant Colony Optimization, Quality of Service, ZRP, Energy -efficiency, Head node.

Received on DD MM YYYY, accepted on DD MM YYYY, published on DD MM YYYY

Copyright © YYYY Author et al., licensed to EAI. This is an open access article distributed under the terms of the Creative Commons
Attribution licence (http://creativecommons.org/licenses/by/3.0/), which permits unlimited use, distribution and reproduction in any
medium so long as the original work is properly cited.

doi: 10.4108/_______________

1. Introduction military battlegrounds for the soldiers to remain connected,

provincial applications [] including conferences etc.,
Ad hoc networks [] have endured several issues [] arising commercial applications [] including disaster management
consequentially due to their unique features [], some of the etc. With increasing interest towards the use of MANETs []
main amongst the others are restricted battery life [], scarce as new applications [] evolve there arises a need for efficient
memory [] and limited processing capabilities []. However, routing protocols that are capable of reducing the constraints
the advantages [] associated with MANETs make them [] characterizing MANETs. Where, the energy efficiency [] is
uniquely popular for certain crucial applications [] such as one of the crucial requirement of the protocol as it will

1Corresponding author. Email:author@emailaddress.com

1
A. N. Author, B. N. Author and C. N. Author

eventually help in increasing the network lifetime, reducing constituting of several artificially generated agents following
the communication overhead [] as well as other limited a structured behaviour without the central administrator, the
resources such as memory [] also becomes one of the critical behaviour is modelled on the basis of real social insect
requirements of the routing protocol []. It is a well-known behaviours with themselves and with the environment. Thus,
fact that ad hoc network faces a restricted energy capacity [] such systems are able to find solutions to problems
owing to its infrastructure-less operation. Thus, a mechanism pertaining to practical relevance.
should be introduced in the ad hoc network operation which The use of the swarm intelligence methodology offers
renders the path selection for forwarding the packets multiple advantages when compared to the classical
incorporating the energy element of the path along with the methodologies. Some of them are:
number of nodes in the path. It is a well proven fact [] that
the lifetime of the network depends on the battery power [] • Adaptability: the group of created swarms are flexible to
available with the nodes, to sustain the network operations. the changes in the environment.
The battery research observes a dearth of significant • Robust nature: In case of failure of one or more individual
development [] and therefore, it becomes crucial to the entire group is still capable of performing the task.
emphasize on the energy management [], as in the absence of • Self organising: the overall group require little
any significant research for improving the battery life of the administration and is majorly capable of self-organisation.
devices, the entire consequence is borne by the routing
protocol as it possesses the central role in the functioning of These attractive properties of swarm intelligence render the
MANET. Energy efficient routing in ad hoc networks approach to be entitled as an eminent design paradigm.
includes discovering paths with minimal energy Natural group activities based on behavioural patterns of
consumption, reducing the dissipation of processor power of several swarms without the controlling of any individual
the computing devices. Mobile devices in ad hoc networks swarm in the group have been investigated. The various
are dependent on restricted power source (which cannot be swarms included in these studies are ants, bees, termites,
generated by the devices on their own) which determines the wasps, etc. The researches have concluded that the ants
longevity of the network. It becomes extremely important for possess the most captivating swarm-level behaviours. The
the protocol to be energy efficient so that the network can be pattern of ants while looking for food is quite fascinating.
sustained for a longer amount of time. The highly mobile They not only follow the shortest route leading to the food
nodes forming an ad-hoc network are powered with the help source but also communicate in a special way (called
of exhaustive resource known as battery. The lifetime of a stigmergy) inorder to share the information with the other
network largely depends on the approaches incorporated ants. Stigmergy is the mode of social networking approach of
within the protocols [12] in-order to reduce the energy indirect coordination used by ants in order to solve foraging
consumption, in the absence of any advancements in the field problems. The realization of stigmergetic coordination is
of battery performance. capable when one ant leaves a trail in the environment by
To ensure a longer battery life, there is a need to incorporate emitting a chemical material known as pheromone that acts
efficient usage of battery energy, which can be achieved as stimulus for the other ant. The observation and in-depth
either by reducing the energy dissipation during the active study of real colonies of ants has led to the development of
state in which the node actively takes part in the ant algorithms and to their application to multiple diverse
communication by either transmitting or receiving packets, optimization problems.
or during the inactive state where the node is idle and listens
to the wireless channel anticipating any routing request from
the neighbouring nodes. Since one of the main apprehensions 2. Framework
in ad hoc network is how to work towards maximising the
lifetime of the network, which can only be possible through 2.1. Route Aggregation
maximum saving of energy, it arises a need for a routing
algorithm capable of incorporating energy constraints. This It works by superseding a group of routes and replacing it
problem of efficient usage of energy can be formulated as an with a lone route for advertisement. This proves effective in
optimization problem and metaheuristics can be employed to curtailing the size of routing tables as the space for a group is
choose optimal route considering the energy level of the now taken by only a single route. The various advantages of
route. RA [11] [14] [15] are:
Over the years, Swarm Intelligence has been the driving
force behind the evolution of multiple algorithms. Nature has a) Rendering the routing process significantly
provided a major stimulus to the development of swarm efficient.
intelligence. This branch of Artificial Intelligence models b) Lessening the load on the processor as well as memory
and designs computational methods based on complex but requirements are reduced.
insightful group behaviour of social insects such as ants, c) The topology changes are isolated.
bees, termites etc. One of the major features related to the
collective behaviour of swarms is self-organization. These
2.2 Clustering
unique features of swarms and their study has led to the
development of swarm intelligence based artificial systems

2
Paper title

Clustering approach is a method which involves splitting the ________________________________________________

entire network into interconnected groups of substructures
known as clusters. Every Cluster elects a coordinator within
itself, responsible for communication in the cluster and is
known as Cluster Head. The advantages of clustering are: 2.3.1 Setting Parameters and initializing
pheromone trails
i. Attempts to strive towards attaining communication
ACO operation is based on the process of simulating a large
scalability in case of an increase in the number of nodes
number of artificial ants in order to model the natural
and high node mobility.
behaviour of real ants. Artificial ants are however different
form the real ants in many ways. Artificial ants possess
ii. It also helps in reducing the number of
memory capabilities which helps them to save the routes
retransmissions and collisions.
followed by them which makes it easier for them to retrace
their return path. Artificial ants deposit the chemical called
iii. It also minimizes the amount of information to be
pheromone only while returning and the memory stored path
stored and propagated within the network.
decides the amount of pheromone to be deposited. The
2.3 Ant Colony Optimization evaporation of pheromone follows an exponential function.

The ant colony behaviour while discovering food source has

enticed a huge amount of interest while dealing with
2.3.2 Generating Solutions
combinatorial optimization issues. Ant Colony Optimization
(ACO) meta-heuristic basis itself on the reverse engineering The artificial ants thus created on each iteration explore the
of the natural ant colony behaviour. A metaheuristic is network to build solutions by visiting unique node each time
formed of two words ‘meta’ meaning of high level and and the selection of next node is based on the pheromone
‘heuristic’ meaning to search. Metaheuristic thus refers to the value associated with that path. The probabilistic rule drives
high-level algorithmic abstraction providing a blueprint to the ant while choosing the solution element based on the
evolve heuristic approaches to find solutions to optimization intensity of deposited pheromone as well as heuristic data.
problems. A Meta-heuristic framework requires lesser During or after the construction of a solution, it is the ant’s
modifications and possess applicability to all general purpose duty to evaluate the solution and determine the amount of
and diverse optimization problems. Amongst the various pheromone to be deposited along the path. The probability
popular metaheuristic algorithms Ant Colony Optimization rule guiding the movement of ant using a stochastic decision
metaheuristics has proved to be one of the most promising. making policy based on pheromone and heuristic information
ACO works by incorporating a colony of artificial ants in as depicted in Equation 1.
order to develop solutions to a particular optimization

{
α β
problem it also incorporates a mechanism of providing a [ τ ij ( t ) ] ∗[ η ij ] k
feedback pertaining to the quality of these solutions by
k α β
, j ϵ Ni
reminiscenting the same communication procedure adopted
by ants in the real world. Artificial ants work as agents and
Pij ( t )= ∑ [ τ il (t ) ] ∗[ ηil ] . (1)
lϵ N ik
are allocated with computational resources and are instilled k
with capabilities of indirect communication with the help of
0 , j ∉ Ni
stigmergy. Stigmergetic communication is based on the
emission of a substance called pheromone laid by the ant
while returning back thus attracting other members to follow
the same path thus reinforcing the route by adding more The term P_ij^k (t) refers to the probability with which ant k
pheromone on the same path. Over a period of time the will drift from node i towards node j during the tth iteration.
pheromone gets evaporated and loses its attractive N_i^k refers to the group of neighbourhood nodes
potentiality hence leading to a solution that the shorter path corresponding to the kth ant at the ith node. The condition
gets visited frequently and thus less pheromone gets when P_ij^k (t) is equal to zero refers to the condition that
evaporated thus providing optimal solution. The pseudo code the ants are forbidden to move to any node outside the
of ACO is detailed as: neighbourhood. [τ_ij (t)]^α refers to the amount of
________________________________________________ pheromone deposited on the path between node i and j with α
Algorithm Ant Colony Optimization Metaheuristic as the weight. [η_ij ]^β refers to the heuristic value on the
________________________________________________ path between i and j with β as weight. η_ij remains constant
while the algorithm executes. The weights α and β control
Set parameters, initialize pheromone trails the level of pheromone and heuristic weights. If α is high
While (not_termination) then it indicates the ants follow the pheromone strictly. If
Generate solutions () low, then algorithm works heuristically.
Pheromone update ()
End while

3
A. N. Author, B. N. Author and C. N. Author

2.3.3 Pheromone Update • To design and develop an efficient scheme for cluster
design as well as an election mechanism for head node
Initially the amount of pheromone trails pertaining to all the with maximum residual energy.
routes are initialized to a minute constant value. Then after • To aggregate all routes at the Head of the Cluster.
constructing a solution path, the pheromone trails undergo • To apply ACO algorithm for energy efficient route
updation in two ways depicted by the following equation. generation.
The equation 5.2 includes the evaporation phenomenon • To implement ACO algorithm for updating the head
determining the amount of the pheromone trails pertaining to node
the paths decreased following an evaporation rate (ρ) that is based on the maximum residual energy and
important for the ants so that they can ignore the suboptimal amalgamating the route aggregation approach to ensure
routes previously converged. The rate of pheromone reduced overhead and longer network lifetime.
evaporation is usually fixed to be appropriately fast so as to
support the investigation of new parts of the search area, and
evade the algorithm from prematurely converge towards
local optimum.

m
τ ij ( t+1 ) ← ( 1−ρ )∗τ ij ( t )+ ∑ ∆ τ ijk ( t ) , ∀ i , j ϵA , 0 ≤ ρ<1
k=1
. (2)

The equation 5.3 involves the increase of the values of

pheromone trails deposited at the visited paths which in
direct proportion with their tour quality. This increase in the
level of pheromone deposited helps implementing a useful
procedure of exploitation of the better-quality paths by
increasing their probability of being utilized yet again by
future ants. The quality paths selected would be including the
solution elements that have been used by a number of ants
previously, or would have been used by atmost one ant and
which resulted in a high-quality solution.

{
Q
k
Δ τ ( t )= C (t ) , if arc ( i , j ) ϵ T k (t)
ij
k
. (3)
0 , otherwise
Where Q is application specific constant
k
C k ( t ) is the overall cost function of the tour T (t) Figure 1. Proposed Algorithm
Figure 1 shows the steps followed in the proposed algorithm.
The several iterations performed in the algorithm render each Each step is elaborated in the following sections.
and every node to be informed of the best neighbours to be
forwarded packets to when looking for a specific destination 3.1 Energy Efficient ACO Algorithm
node. However, this algorithm can be modified inorder to
optimize the memory utilization and also incorporate the
energy element of the path into the algorithm and one such
modification is explained in the next section.

3. METHODOLOGY
The elemental objective of the research paper is to proffer a
highly reliable and dominantly efficient MANET routing
protocol which is capable of performing energy efficient
route aggregation with the help of ACO metaheuristics. With
a view to achieve the desired aim, the research
comprehended the following objectives

4
Paper title

Considering the memory record inorder to search the ant

identification. In case of not finding the record, saving the
required data. The node restarts a timeout timer for the ant.
To avoid any loop, the node eliminates the ant if a similar
record is found in the memory. On receiving a backward ant,
the node explores its memory and locates for the next node
towards which the ant must be forwarded. The timeout timer
helps when to eliminate the record identifying the backward
ant if the timer times out.

3.1 ACO aided Cluster Design and Head Selection

Figure 1. Algorithm

Considering the memory record inorder to search the ant

identification. In case of not finding the record, saving the
required data. The node restarts a timeout timer for the ant.
To avoid any loop, the node eliminates the ant if a similar
record is found in the memory. On receiving a backward ant,
the node explores its memory and locates for the next node
towards which the ant must be forwarded. The timeout timer
helps when to eliminate the record identifying the backward
ant if the timer times out.
The EEABR incorporated few improvements to the
conventional ant colony optimization algorithm with a view
to reduce the memory utilization in the network nodes and
also incorporated a consideration for the energy quality
related to the paths explored by the ants. 3.2 AC
The conventional algorithm includes sending of the the O
forward ants irrespective of the destination node, thus
requiring communication between the nodes which leads to
the storage of identification pertaining to the neighbouring
nodes as well as their corresponding levels of pheromone
trail into the routing tables of the nodes. However, it
becomes a problem for the larger networks, owing to the
amount of memory needed to store such information.
To save memory utilization in WSNs, the forward ants can
be sent straight towards the sink-node, thus reducing the
memory to be used only for the neighbours in the same
direction as the sink node. Thus, considerably reducing the Aided Route Aggregation
size of the routing tables consequently, the memory required.
The improved algorithm in attempt to reduce memory ALGORITHM FOR ACO AIDED ROUTE
reduces each ant’s memory Mi to just couple of records AGGREGATION
which is only the two latest visited nodes. The path taken by
the ants in this algorithm is stores in the memory at each 1. For
3.3 ACOeachAided
Cluster Head -efficient
Energy in the cluster.
path Selection
node which now records the details of each ant. A memory 2. Apply Route Aggregation (using
record is created at each node which stores the previous node TheACO)
ants help us to determine routes between the nodes with
visited by the ant, the next node to which the ant was 3. Updateresidual
maximum Routingenergy.
Tables Amongst
within thethe multiple routes
forwarded, identification number of the ant and an associated Cluster.by the ants, the routes with maximum residual
discovered
timeout value. The process followed by the node after 4. Maintain
energy andare
of nodes update
chosenthefor
cluster
data transmission. The
receiving a forward ant is as follows: Head used
algorithm with highest
for routeresidual energy. for data transmission
determination
is5.exhibited
Move toinnext Cluster Head.
Fig..

5
 This is the title

4.RESULTS
The proposed mechanism enables the perfect
exploitation of residual energy of the node and prevents
the premature dying of the node led by energy starvation
and thus improving the network lifetime. The energy
efficient ZRP-RA achieves an enhancement by further Average Energy Consumption
applying Route Aggregation at the Cluster Head. Table
Number of
details the simulation parameters while simulating the
Nodes 25 50 75 100
proposed approach using NS2.
ZRP 0.69 0.62 0.67 0.72
Simulation Value ACO-ZRP- 0.61 0.53 0.57 0.63
RA
Parameters
Network Area 1000x1000 m2 Table reveals that Ant Colony Optimization ensures a
marked decrease in the energy consumption as compared
Velocity 1~2m/s to ZRP. Thus, accomplishing performance enhancement
No. of Nodes 25,50,75,100 as compared to the conventional hybrid protocol.
Packet Size 512 byte
Traffic Type CBR
Number of Connection 20
Packet rate 2 P/s
Pause Time 0
Simulation Time 100 sec

Packet Delivery Ratio

Number of
Nodes 25 50 75 100
ZRP 42.21 98.38 51.41 35
ACO-ZRP- 71.78 99.82 76.28 63.71
RA

Throughput
Number of
Nodes 25 50 75 100
ZRP 46.3 106.15 56.23 38.02
ACO-ZRP- 97.37 126.28 103.72 81.67
RA

EAI Endorsed Transactions

6 on ......................................
.................-.............. 2013 | Volume .... |
Issue ....-.... | e...
 14.989% reduction in average jitter value,
Throughput improvement has increased by
49.22%, the average energy consumption has
been reduced by 9.2835% and Packet loss ratio
is reduced by an impressive 51.61% hence, when
scalability is increased average delay and jitter of
the proposed approach does not rise too heavily,
delivery ratio is also maintained and throughput
also does not fall down too heavily. Therefore, it
can be specified the proposed methodology is
quite effective with respect to the proposed ZRP-
RA approach of mobile ad-hoc networks as
various QoS parameters like delivery ratio,
average delay, jitter and average throughput are
maintained to provide an enhanced energy
efficient optimized solution for MANET.

A.1. The first subappendix

This is the body text with no indent. This is the
body text with no indent. This is the body text
with no indent. This is the body text with no
indent. This is the body text with no indent. This
5. DISCUSSION is the body text with no indent.
ACO technique provides the energy efficient
optimal paths for data transmission and Route . (A.1)
Aggregation mechanism efficiently reduces the
overhead by aggregating the information at the This is the body text with no indent. This is the
head hence reducing the energy consumed even body text with no indent. This is the body text
in high density conditions. It also reduces the with no indent. This is the body text with no
queuing of packets by reducing congestion in the indent. This is the body text with no indent. This
network as a result of efficient aggregation of is the body text with no indent. This is the body
information at the head along with the circulation text with no indent. This is the body text with no
of the cluster head as the energy level decreases indent.
even in high density conditions thus, ensuring an This is the body text with indent. This is the
improvement in the amount of packets delivered body text with indent. This is the body text with
which consequently results in an improved indent. This is the body text with indent. This is
throughput and thus achieves further the body text with indent.
enhancement in the performance and produce
better results.

6. CONCLUSION
Meta Heuristic Algorithm based ACO
approach is utilized for further energy efficient
enhancement of the proposed ZRP-RA protocol.
The importance of the ACO applied ZRP-RA
protocol for MANET is further specified by
simulating it and comparing it with the ZRP-RA
approach. 30.62% increase in packet delivery
ratio, reduction in delay value by 23.96%, a
 This is the body text with indent. This is the
body text with indent. This is the body text with
indent. This is the body text with indent. This is
the body text with indent. This is the body text
with indent. This is the body text with indent.

Appendix B. The second appendix

This is the body text with no indent. This is the
body text with no indent. This is the body text
with no indent. This is the body text with no
indent. This is the body text with no indent. This
is the body text with no indent. This is the body
text with no indent. This is the body text with no
indent. This is the body text with no indent. This
is the body text with no indent. This is the body
text with no indent.

Acknowledgements.
This is the acknowledgement text. This is the
acknowledgement text. This is the acknowledgement
text. This is the acknowledgement text. This is the
acknowledgement text. This is the acknowledgement
text.

References
[1] Journal article: SURNAME, A. (Year) Title of
paper. Title of Journal volume(issue): page
numbers.
[2] Book: SURNAME, A. and SURNAME, B. (Year)
Title of Book, 2nd ed. (Publisher location:
Publisher).
[3] Chapter: SURNAME, A. (2010) Title of chapter. In
SURNAME, C. [ed.], Title of Book (Publisher
location: Publisher), ch. 3.
[4] Conference: SURNAME, A. (2010) Title of paper.
In Proceedings of Title of Conference,
conference location, conference date (Publisher
location: Publisher), xxx–yyy.