Journal Pre-proof

Energy Efficient Cluster based Routing Protocol for WSN using

Butterfly Optimization Algorithm and Ant Colony Optimization

Prachi Maheshwari Dr. , Ajay K. Sharma , Karan Verma

PII: S1570-8705(20)30673-9
DOI: https://doi.org/10.1016/j.adhoc.2020.102317
Reference: ADHOC 102317

To appear in: Ad Hoc Networks

Received date: 8 May 2020

Revised date: 24 August 2020
Accepted date: 2 October 2020

Please cite this article as: Prachi Maheshwari Dr. , Ajay K. Sharma , Karan Verma , Energy Efficient
Cluster based Routing Protocol for WSN using Butterfly Optimization Algorithm and Ant Colony Opti-
mization, Ad Hoc Networks (2020), doi: https://doi.org/10.1016/j.adhoc.2020.102317

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier B.V.

Energy Efficient Cluster based Routing Protocol for WSN using Butterfly Optimization Algorithm
and Ant Colony Optimization

Prachi Maheshwari1*
National Institute of Technology, Delhi
[email protected]/ [email protected]

Ajay K. Sharma2
IKG Punjab Technical University, Jalandhar, Punjab, India
[email protected]

Karan Verma3
National Institute of Technology, Delhi
[email protected]
Abstract

Wireless Sensor Networks (WSNs) consist of a large number of spatially distributed sensor nodes
connected through the wireless medium to monitor and record the physical information from the
environment. The nodes of WSN are battery powered, so after a certain period it loose entire energy. This
energy constraint affects the lifetime of the network. The objective of this study is to minimize the overall
energy consumption and to maximize the network lifetime. At present, clustering and routing algorithms
are widely used in WSNs to enhance the network lifetime. In this study, the Butterfly Optimization
Algorithm (BOA) is employed to choose an optimal cluster head from a group of nodes. The cluster head
selection is optimized by the residual energy of the nodes, distance to the neighbors, distance to the base
station, node degree and node centrality. The route between the cluster head and the base station is
identified by using Ant Colony Optimization (ACO), it selects the optimal route based on the distance,
residual energy and node degree. The performance measures of this proposed methodology are analyzed
in terms of alive nodes, dead nodes, energy consumption and data packets received by the BS. The
outputs of the proposed methodology are compared with traditional approaches LEACH, DEEC and
compared with some existing methods FUCHAR, CRHS, BERA, CPSO, ALOC and FLION. For
example, the alive nodes of the proposed methodology are 200 at 1500 iterations which is higher
compared to the CRHS and BERA methods.

Keywords: Ant colony optimization; Butterfly optimization algorithm; Energy consumption; Network
lifetime; Wireless sensor networks.
 1. Introduction

WSN contains hundreds or thousands of sensor nodes, which are distributed in the environment
to sense, evaluate and receive data. These sensor nodes are inexpensive and have higher capacities in
sensing, processing, and transmitting information [1]. WSN is used in several applications like weather
forecasting, defense domain, medical field and various commercial and industrial applications [2, 3]. The
sensors of WSN are compact and uses a limited battery as their power source [4]. The sensor collects data
using an Analogue to Digital Converter (ADC) and it processes the data for transmitting to the main
location that is known as Base Station (BS). At BS, the received data are analyzed to make decision for
various applications. The sensor nodes of WSN operate as a repeater to transmit information to other
sensors and the sink [5, 6, 7]. Moreover, the WSN‟s power source should be used in an adequate way as it
cannot be exchanged or recharged due to the sensor placement at harsh and no-man environment [8, 9].
The architecture of WSN affected by various parameters like scalability, fault tolerance, energy
efficiency, etc. [10, 11]. The sensors present in WSN exhaust its energy in two ways i.e., 1) Sensing of
environmental parameters and 2) Data transmission to BS through the nodes. The data transmission in
WSN utilizes higher energy than data sensing and processing from environment [12, 13].

The insufficient power sources of the sensor nodes are considered as a key issue in WSN.
Consequently, the failure in network arises because of node failure [14, 15, 16, 17]. Therefore, the major
difference between WSN and other standard wireless networks is that WSNs are hypersensitive and
susceptible to the energy [18]. The sensor nodes exhaust its energy quickly because of direct data
transmission from each sensor to BS [19]. Moreover, the optimal energy utilization in WSN is required to
achieve higher lifetime and improve performance of WSN [20]. Therefore, grouping of sensors into
clusters is utilized to decrease consumption of the network energy and increase the scalability of the
network [21, 22, 23]. Each cluster of a network has one header, known as Cluster Head (CH) that contacts
with other CHs of the network [24]. As, higher amount of energy is required to directly transfer the
sensed data to the BS, a routing protocol is used in the clustered WSNs to identify the best route between
the CHs and BS to reduce energy consumption [25, 26]. The features of the routing protocols include
fault tolerance, reliability, information accumulation, scalability, etc. [27].

The objective of this study is to decrease energy consumption of the nodes during data
transmission. This helps to increase the total packet transmission to BS by decreasing the energy
consumption by the sensor nodes. The swarm intelligence is mainly used in this research due to its
searching capacity, robustness and self-adaptability.
 The main contributions of this research are stated below:

 The BOA is used for selecting the CH in the WSN due to its high stability and less computational
complexity. In this study, BOA selects the CH based on multiple objective value residual energy,
distance to the neighbors, distance to the BS, node degree and node centrality.
 The shortest path between CH to BS is identified using the ACO, because the ACO has capacity
to deliver rapid discovery of solutions in WSN. Here, the ACO is optimized with residual energy,
distance and node degree to overcome the limitation of the uncertain convergence time.
 Lifetime of the network is increased because of the energy efficient CH selection and optimal
route generation for data transmission. In addition, total packets received by the BS are increased
by minimizing the energy consumption of the nodes while transmitting the data packets.

The paper is organized as follows: The literature survey of recent researches related to the
clustering and routing techniques is presented in Section 2. The issues found from existing researches
along with the solution is discussed in Section 3. The different preliminaries considered in this study are
presented in Section 4. A detailed explanation about the cluster based routing using BOA and ACO
algorithms is presented in Section 5. The performance and comparative analysis of the proposed
methodology with existing algorithms are discussed in section 6. At last, the conclusion is drawn in
Section 7.

2. Literature survey

There have been various researches developed to optimize WSN performances such as energy
cost, network lifetime, etc. The minimization of sensor node‟s energy consumption is considered as one
of the most important features in WSN. Therefore, clustering technology has been used extensively to
handle deficiency of the sensor network. A conventional network routing protocol is inefficient due to
variations in the power supply of energy constrained wireless sensor networks. The researchers have
concentrated more on clustering and routing techniques. There are various researches have been made in
classical approaches [28-32], swarm intelligence based approaches [33-43] and especially ACO based
swarm approaches [44-45] from these, some researches are discussed in this section.

2.1. Classical approaches

Wang et al. [28] presented an improved LEACH (LEACH-Impt) protocol to enhance the
performance of the traditional LEACH protocol. Several disjoint paths were created in LEACH-Impt
based on the routing topology of inter cluster nodes. Moreover, the route hops, path residual energy and
energy consumption were considered during the selection of an optimal path. The energy consumption of
the LEACH-Impt was less compared to the standard LEACH protocol. But, the limitation of the LEACH
protocol is a random selection of CH from a set of nodes. The information collection and transmission are
lost due to changes in network topology.

Haseeb et al. [29] introduced a Weighted Energy-Efficient Clustering with Robust Routing
(WECRR) protocol, which selects the CHs and divides the network into clusters. The routing decision
was made by considering various factors like traffic density, residual energy and packet error ratio. In
addition, WECRR improved route maintenance and load distribution. The route maintenance was
initialized to decrease unwanted energy consumption. The energy balancing between the nodes was
obtained using WECCR. But, the residual energy of the nodes is considered as higher priority compared
to the distance between the node and BS.

Haseeb et al. [30] presented an Aware Cluster Based Routing (AECR) protocol. This AECR
protocol generated distinct clusters based on node density and positions. It generated the balance sized
clusters by considering the distribution of the sensors and it eliminated the formation of random clusters.
The intra-cluster route discovery was achieved by broadcasting the HELLO packet among cluster
members. The clustering overhead and communication cost were reduced by mitigating frequent re-
clusters formation. In this formation of AECR protocol, the energy of the node is considered as higher
priority when compared with the distance between the nodes.

Haseeb et al. [31] presented dynamic energy-aware fault tolerant routing protocol. Initially, the
network was partitioned into uniform sized clusters based on network size. The nodes were selected as
CH using compound weighted metrics. This fault aware routing created the parallel routing chains to
ensure reliable data forwarding path between nodes. The distance, residual energy and link quality were
considered in the next-hop neighbor selection to balance the energy consumption. In this study, an
optimum number of cluster identification is required otherwise it leads to insufficient use of energy and
computational resources.

Han and Zhang [32] developed an energy efficient clustering routing protocol named as WPO-
EECRP. In WPO-EECRP protocol, multiple clustering factors were considered for CH selection, such as
residual energy, distance from the node to the BS and number of neighbor nodes. This WPO-EECRP
protocol had good scalability and it also provided an adequate clustering control by altering the clustering
parameters. The direct data transmission from CH to BS happened, when CH didn‟t had any adjacent
nodes. These types of communication increase the energy consumption through the network.

2.2. Swarm intelligence based approaches

Morsy et al. [33] presented a Gravitational Search Algorithm (GSA) for finding optimal CHs
from the group of nodes. The node‟s residual energy was considered to choose the CH. The objective
functions considered in CH selection were energy efficiency, intra cluster distance and distance to BS.
One hop approach was used for data transmission between the sensor node and CH. Subsequently, multi
hop approach was used to select next hop based on the cost function to transfer the data over WSN. The
cost function of the multi hop approach considers only the residual energy of the next hop and its distance
from the BS.

Lalwani et al. [34] presented a Harmony search algorithm (HSA) to cluster the network and
create the route in WSN. The fitness functions considered in clustering are node degree, energy, intra-
cluster distance and distance between the CH and BS. In addition, the fitness functions considered in the
HSA routing were node degree, energy and distance. The energy consumption of the node was reduced by
selecting the node with minimum distance for data transmission. The performance of the network affects
when the clusters have a number of non-CH members.

Lalwani et al. [35] proposed an Optics inspired optimization (OIO) for both clustering and
routing in WSN. The clustering and routing was optimized using three different parameters node degree,
energy and distance to achieve optimal clusters and transmission path, respectively. In clustering process,
the OIO selected the CH first, then potential function was initiated to assign the nodes to the desired CH.
The next hop node (i.e. CH) with more energy was selected to avoid the packet loss in the network. The
computational complexity of the OIO based clustering and routing increase exponentially with the
network size.

Lalwani et al. [36] presented a Biogeography Based Optimization (BBO) to select CH from the
network by considering fitness function values like distance among CHs and the BS, distance between the
nodes of cluster and residual energy. The BBO was used to generate the optimal shortest route between
the CH and the BS. The data packets received by BS were maximized by preserving the residual energy
of the sensors. The remaining energy of the network was slightly degraded compared to the genetic
algorithm based routing when the BS was located outside the network.

Rao and Banka [37] presented a novel Chemical Reaction Optimization (nCRO) to obtain an
unequal clusters and generate the transmission path in WSN. In CH selection, the nCRO was used with a
potential energy function and molecular structure encoding to select the CHs based on the average sink
distance, energy ratio and node distance. Next, the normal sensor nodes were allocated to the desired CHs
based on the cost function. The potential energy function of nCRO was used to identify the route from
source to the BS. The unbalanced energy consumption was reduced by selecting the optimal CHs from
the network. The fault tolerance is not considered in this nCRO based clustering and routing algorithm.

Mekonnen and Rao [38] developed the Simulated Annealing (SA) and Particle Swarm
Optimization (PSO) based polynomial time clustering algorithms in WSN namely CSA and CPSO
respectively. These clustering algorithms consisted of two phase setup and steady state phases. The CH
selection and cluster formation were accomplished in the cluster setup phase. The modes of normal nodes
were fixed into sleep mode until it was used to perform data transmission. Next, the normal nodes
transmitted the data to its respective CH within their own Time Division Multiple Access (TDMA) time
slot. Moreover, CH collects the data packets from its own normal nodes and transfer to the BS.

Yogarajan and Revathi [39] presented Ant Lion Optimization for Clustering (ALOC) technique to
enhance the energy efficiency of the network. The fitness function used in the ALOC considered the
residual energy, amount of adjacent nodes for each node, distance between the nodes and distance
between the node and the BS. The node with high amount of fitness was considered as CH. Then the
optimal route between the CHs to the BS was obtained using discrete ant lion optimization algorithm.

Ezhilarasi and Krishnaveni [40] developed the Evolutionary Multipath Energy-Efficient Routing
(EMEER) protocol for WSN. In EMEER protocol, the cuckoo search algorithm clustered the network
based on the similarity and energy-level features. Then CH was selected based on the distance from the
node to sink, energy and trust value. The energy consumption of the network was less when compared
with the standard algorithms LEACH, PEGASIS and TEEN protocols. This EMEER protocol fails to
analyze the alive node performances during communication.

Gao et al. [41] presented PSO based energy constrained clustering routing to minimize the energy
consumption of the WSN. The sensor node with higher residual energy was considered as candidate CH
during the clustering process. After completing the PSO clustering, the process of data transmission was
initialized between the nodes to the CH and CH to the BS. The data collection, aggregation and
distribution of CH was required an additional energy in the WSN.
 Sirdeshpande and Udupi [42] developed the Fractional lion (FLION) clustering algorithm to generate
optimal routing path in the WSN. The fitness function of the FLION algorithm was considered 5
objectives such as inter cluster distance, intra cluster distance, energy of CH and normal nodes and delay.
The neighbor solution was detected by incorporating the fractional order derivative in the FLION
algorithm. The rapid CH selection using FLION algorithm was used to improve the network lifetime.
This FLION clustering algorithm doesn‟t consider the node degree during clustering process.

2.2.1. ACO based swarm approaches

Xie et al. [44] presented a combination of Type-2 Mamdnai Fuzzy Logic System (T2MFLS) and
Ant Colony Optimization (ACO) for WSN. The CHs were selected using the T2MFLS which considered
the residual energy, neighbor nodes and distance to BS. Subsequently, ACO was used to identify a
routing path to connect all CHs. Each CH communicated with neighbor nodes and it transmitted data to
the BS. The disconnection of the WSN was avoided by considering the residual energy of each node in
the cluster technique. The density of the nodes affects the performance of the network.

Arjunan and Sujatha [45] presented a fuzzy logic to produce an unequal clustering and ACO for
routing between the nodes. The nodes were clustered and the CHs were selected using fuzzy logic that
considered the residual energy, distance to BS, distance to its neighbors, node degree and node centrality.
The ACO based routing was used to obtain route from CHs to BS. The cluster maintenance phase was
used in the WSN which made it to consume an equal energy by all CHs. This helps to improve the
lifetime of the WSN. The number of transmissions is less, because the implemented method is operated in
a both proactive and reactive manner.

An effective technique is required to be implemented to preserve the node‟s energy level while
transmitting the data from the source to the destination. So, an energy efficient WSN is need to be
developed because of the limited energy of the sensor nodes present in the WSN. In WSN, the limited
energy of sensors makes it expensive and difficult to obtain better network performances. Therefore, the
self-configuring approach is developed in the energy aware routing to improve the network performances.
The major concerns of the WSN are data routing and transmission to the BS over the network. So, an
efficient routing with optimal path selection and optimal energy consumption is developed to identify the
optimum path from the source node to the BS.

3. Problem statement

Current issues of WSN are discussed in this section and it provided how the proposed
methodology gives the solution to this problem. The problems of present WSN are:

An appropriate selection of objective functions need to be considered in the network for creating
energy efficiency in WSNs. Similarly, WECRR [29] and AECR [30] gave higher priority to the residual
energy of the nodes in the clustering. The energy consumption of the network reduces only when the
respective technique gives equal priority to both energy and distance. The energy efficient WSN should
work in both small scale and large scale WSN applications. In HSA based clustering and routing [34], the
performance of WSN was affected when CH had a number of non-CH members in its cluster. Similarly,
the density of the nodes affects the network performance of CRT2FLACO [44]. The direct data
transmission among CH to BS consumes more energy in WSN [32]. It gives rise to the hot spot problem
between networks that creates packet loss through the network. Moreover, the data transmission through
ACO based routing [45] creates packet loss between networks due to the combination of proactive and
reactive algorithms. The sensor node becomes unreliable and faulty, due to the deployment of nodes in
the harsh and unattended environment. Moreover, the energy consumption of the nodes is also considered
as main issue while transmitting the data packets to the destination. Because, the nodes with in sufficient
energy causes a packet drop in data transmission.

Solution:

In this study, both energy and distance are considered to create energy efficient WSN because the
energy consumption of each node mainly depends on the distance between the nodes, so energy
consumption is directly proportional to the distance between the nodes. In addition, the packet loss is
avoided by considering the energy of each node in WSN. The problem in routing is avoided by
implementing the multi hop routing over the network. ACO technique is used to create the route from the
source to BS. The objective functions considered in the routing are residual energy of the nodes, distance
among the nodes and number of hops in each cluster. By considering these parameters, the packet loss
between networks is reduced. Therefore, an energy efficient WSN is designed to work in both small and
large scale WSNs effectively.

4. Preliminaries

This section provides information about the network model, energy model, description of the butterfly
optimization and ant colony optimization algorithm.
 4.1. Network model

The structure of WSN is presented in figure 1. The network model is formulated based on the
following considerations:

 In WSN, all the sensor nodes are similar to each other in terms of initial energy and processing
time.
 The distance between the sensors is calculated based on the formula of Euclidean distance.
 The sensor nodes are randomly deployed in the sensing area and the position of the sensors is
constant after the deployment.
 BS receives the information about the residual energy and distance from the sensor nodes. Based
on this information, CHs are selected for all the sensor nodes by using an effective CH selection
algorithm. Then, the routing process is used to obtain the path between the CHs to the BS.

Figure 1. Structure of cluster based WSN

 4.2. Energy model

In this study, a basic 1st order radio model is considered to calculate the transmitter and receiver
energy. The consumed energy to transmit and collect the bit packets over the distance are expressed in
equations (1) and (2) respectively.

( ) { (1)

( ) (2)

where, the amount of energy dissipated at the transmitter/receiver is represented as and the threshold
distance is . The threshold distance is calculated using the following equation (3).

√ (3)

where, the amplification energy for free space and multipath model is represented as and
respectively. These and depend on the transmitter amplifier model.

5. Proposed method

The proposed method has two algorithms: one is used to choose the CH and other one is for
routing over the network. BOA is used to select optimal sensors as CH and ACO select the optimal route
between the CH and the BS. Next, the CHs transmit the collected data to the BS through the generated
path from ACO. The proposed method is broadly described in the following section. The overall process
of the proposed method is depicted in the following figure 2.
 Start

Initialization of parameters

Sensor node deployment in the test area

Cluster head selection using BOA

Fitness function estimation of BOA using residual energy,

distance between the nodes, distance between the CH to
BS, node degree and node centrality

Routing among the CHs using ACO

Fitness function estimation of ACO routing using residual

energy, distance from the CH to the BS and node degree

If Numnodes= No
Number of dead nodes

Yes

Performance measure (Alive nodes, average energy

consumption, throughput, total packets received by the BS,
routing overhead, packet drop ratio and network lifetime)

Stop

Figure 2. Flowchart of the proposed methodology

5.1. Cluster head selection using BOA

Butterfly Optimization Algorithm (BOA) is one of the nature inspired metaheuristic algorithm
developed by Sankalap Arora [46]. This BOA imitates the butterfly‟s food search and its mating behavior.
The behavior of butterflies is characterized based on the search algorithm. Generally, the butterflies are
attracted to each other based on the emitted fragrance of other butterflies. Subsequently, the movement of
butterflies is either in a random manner or in the direction of the butterfly which emits higher fragrance.
Further, the objective function is used to determine the butterfly‟s stimulus intensity. The BOA chooses
the optimal CHs between all sensors by using node degree, node centrality, distance to its neighbors,
distance to the BS and residual energy.

5.1.1. Butterfly representation and Initialization

In CH selection phase of BOA, the butterfly specifies a group of sensors to choose as CHs from
the sensors of a network. Each butterflies dimension is equivalent to the amount of CHs in the network.
The position of each butterfly is initialized with a random node ID among and , where is the total
amount of nodes in the network. Assume, ( ( ) ( ) ( ) ) be the butterfly and
where every butterfly position that specifies the node ID among 1 to in the network
and defines the amount of CHs in the network.

For example, assume the network has 50 sensor nodes and CH will be 10% of the total amount of
nodes. The dimension of each butterfly of BOA is equivalent to the amount of CHs that is 5. Then the
position of each butterfly is that is initialized between the random number from 1 to 50
(i.e., total amount of sensor nodes). The position of each butterfly is mapped along with its coordinates of
sensor nodes.

5.1.2. Butterfly position improvisation

The position improvisation of the butterflies is considered in the process of CH selection. The
process of updating the butterfly‟s positions is given as follows:

a. Generation of new butterflies

In BOA, the positions of the butterflies are updated either in a global search phase or in local
search phase. The selection of global or local search phase is determined by the random number in [0, 1].
The global and local search phase of the BOA are specified in equations (4) and (5) respectively. In that,
new butterflies are randomly improvised with node ID from the entire nodes of the network. The
fragrance emitted by the remaining butterflies improves the butterfly‟s position. In this case, the positions
are updated based on the fragrance of the butterflies. Moreover, the local search is initialized in the BOA,
when the butterfly not able to find any fragrance from the environment.

The butterfly moves based on the best butterfly or solution at global search phase.

( ) (4)

where, the solution vector for butterfly at iteration is ; the random number generated
between 0 and 1 is ; the detected best solution is determined between all solutions of current iteration is
denoted as and the butterfly fragrance is .

( ) (5)
where and butterfly is represented as and . The equation (5) acts as local random walk when
both and are from the same butterfly.

The fragrance of the butterflies is generated in the respective positions by using the following
equation (6).

(6)

where, is the fragrance; is sensory modality; is stimulus intensity and is the power exponent.

The fitness function is derived in the following section 5.1.3 that is used for the computation of
butterfly fragrance. The fitness function is used as the stimulus intensity of the butterfly and then the
fragrance of the butterfly is calculated to update its positions.

5.1.3. Fitness function of BOA

The fitness function of BOA is used to select the optimal CH from the group of sensors in the
network. The residual energy considered in the fitness function is used to avoid the dead node as a CH
during clustering process. Next, the distance between the nodes and the distance from the candidate CH to
the BS are used to select the optimal CH to minimize the energy consumption of the nodes. The node
degree is considered for selecting the CH with less number of normal nodes to preserve the node for
higher rounds. Moreover, the higher centrality to its cluster members leads to minimize the transmission
distance between the cluster members to CH. The considered fitness functions are described as follows:

a. Residual energy of the CH

In a network, CH performs various tasks collecting the data from normal sensor nodes and data
transmission to BS. The CH requires high energy to accomplish aforementioned tasks, so the node with
higher residual energy is preferred to be a CH. The residual energy ( ) is described in the following
equation (7).

∑ (7)

where, the residual energy of the cluster head is .

b. Distance between the sensor nodes

It defines the distance between the normal sensor nodes and its own CH. The node‟s energy
dissipation mainly depends on the distance of the transmission path that is mentioned in section 4.2. The
energy consumption of the node is small when the selected node has less transmission distance towards
BS. The distance from the normal sensors to CH ( ) is expressed in the equation (8).

∑ (∑ ( ) ) (8)

where, the distance between sensor and CH is denoted as ( ) and the number of sensor nodes
that belongs to CH is denoted as .

c. Distance among the cluster head and BS

 It defines the distance from CH to BS. The node‟s energy consumption depends on the distance
through the transmission path. For instance, if BS is positioned far from CH, then it needs more energy
for data transmission. So, the sudden fall of CH may occur due to higher energy consumption. Hence, the
node with a lesser distance from BS is preferred during data transmission. The following equation (9)
represents the objective function of distance between the cluster head and the BS ( ).

∑ ( ) (9)

where, the distance between the CH and BS is denoted as ( ).

d. Node degree

It defines the number of sensor nodes belongs to the respective CH. The CH with less number of
sensors are selected, because the CHs with higher cluster members losses their energy in lesser duration.
The node degree ( ) is expressed in equation (10).
∑ (10)

where Ii are number of sensor nodes belonging to .

e. Node centrality

Node centrality ( ) defines how much a node is centrally located from the neighbor nodes and it
is expressed in equation (11).

√(∑ ( )) ()
∑ (11)

where ( ) is number of neighboring node of .

The weight value is allocated for each objective value. In this case, the multiple objectives are
changed into single objective function. The weighted values are and . The single
objective function is shown in the equation (12).

where, ∑ ( ) (12)

where, the values of and are 0.35, 0.25, 0.2, 0.1 and 0.1 respectively. The is
considered for taking a residual energy as a higher priority to avoid the node failure as a CH.
Subsequently, the and are considered for second and third priority to identify the CH from the BS
with the less distance which leads to minimize the energy dissipation. The fourth priority ( ) is given to
the node degree to select the CH with a less node degree. Additionally, the node centrality is considered
as last priority which used to increase the closeness between the CH and cluster members.

The aforementioned steps are the complete process of CH selection using BOA and its Pseudo code is
given in the “Algorithm 1”
Algorithm 1: Cluster head selection using BOA

1. Fitness function derivation using equation (12).

2. Create an initial population of butterflies.
3. Initialize stimulus intensity ( ) at which is calculated by ( ).
4. Initialize switch probability, sensor modality and power exponent.
5. for j= maximum number of iterations
6. for each butterfly in population
7. Compute fragrance using equation (6).
8. end for
9. Detect the optimal butterfly population
10. for each butterfly in population
11. Create a random number from [0,1]
12. If
13. Move towards an optimal solution using equation (4).
14. else
15. Move randomly using equation (5).
16. end if
17. end for
18. Update the value of power exponent
19. end for
20. Output the CH from the network (Optimal solution).

5.1.4. Cluster formulation using potential function

After selecting the CHs by BOA, the sensor nodes are allocated to the CHs using potential
function mentioned in the equation (13). The sensor nodes are allocated to the CH with less transmission
distance and higher residual energy. Therefore, the amount of consumed energy will be less during the
data transmission phase.
( )
(13)
( )

where, sensor node potential is ; the proportionality constant is and ( ) specifies the
distance among the cluster head and sensor ; ( ) represents the residual energy of the
respective CH; The sensor is allocated to particular CH which has higher potential. If the distance
between the sensor nodes and two different CHs are same then the sensor node connect with the CH has
higher energy.
 5.2. Routing algorithm using ACO

Ant Colony Optimization (ACO) is also a metaheuristic algorithm inspired based on the behavior
of the ants. Generally, the ants discover the shortest path in their nest and also searches for a food source.
ACO is applied to discrete problems, which is demonstrated as a graph with nodes and number of
links. In this initialization, every node has an amount of ants and each link is related to the weight. At
first, the weight of the links is calculated based on the physical distance, generated random number or a
value that is found using the mathematical formula. The limitation of the uncertain convergence time is
overcome by optimizing the ACO with residual energy, distance to the BS and node degree. The route
generation process using ACO [47] is described clearly in this section.

1. An ant is placed in each CH to generate the route from the CH to BS. The source CH creates some
route setup packets that are referred as Forward Ant Packets.
2. The Forward Ant Packets are randomly transmitted to the next CH node based on the probability
matrix. This transmission is continued until the forward ant packets are received by the BS.
3. In this Forward Ant Packets transmission, each packet generates the local database which has the
information about the visited CH (node ID), residual energy( ), distance to the BS ( ) and
node degree( ). The residual energy of the CH is mainly depends on the amount of packets ( )
transmitted through the network.
4. The database of the forward ant is used to create the backward ant packet, when the path is extended
until it reaches BS. Subsequently, this backward ant transmits in the same path which is transmitted
during forward ant packet transmission.
5. The pheromone value is updated for each path based on the residual energy, distance to the BS and
node degree.
6. The ant selects the next hop based on the node transition rule expressed in equation (14) which
expresses the probability of choosing the node as the next node of using ant .

[ ( )] [ ]
( ) { ∑ [ ( )] [ ] (14)

where, the heuristic value and pheromone intensity are represented as and respectively. The
parameters and are used for controlling the relative importance of heuristic value and pheromone
intensity. represents the group of nodes that ant has not visited yet.

The heuristic information and pheromone intensity are updated based on the CH information
stored in the routing table. The heuristic information is updated based on the distance between the CHs
and it is mentioned in the following equation (15).

(15)

where, is the distance among the CHs.

The pheromone update rule is expressed in equation (16).

( ) ∑ (16)
where, the amount of ants initialized in ACO is and the pheromone decay coefficient is . The
quantity of pheromone present in the links and is represented as that is expressed in the following
equation (17).

( )
{ (17)

where, the constant value is and the detected cost of the path by ant is .

The weight value is considered for each fitness value which is used to convert the fitness function
values into single objective namely route cost. The route cost ( ) of equation (17) is updated in the
quantity of pheromone value. The route cost is given in the following equation (18).

(18)

where, the weighted values of and are 0.5, 0.3 and 0.2 respectively. The residual energy is
considered as the first priority to avoid the sensor node with inadequate energy. Because the sensor node
with lesser energy creates node failure during the communication. The distance between the CH and BS is
considered as second priority to obtain the shortest path that used to minimize the energy consumption.
Next, the node degree is considered as third priority to select the next hop CH node with less number of
cluster members.

The aforementioned steps are the process of ACO routing between the source node to the BS and its
Pseudo code is given in the “Algorithm 2”
Algorithm 2: Routing using ACO

1. Initialize pheromone exponential weight and heuristic exponential weight.

2. Create an initial population of ants.
3. for j= maximum number of iterations
4. for each ant in the population
5. Repeat until ant completes the tour to the destination.
6. Select the successive ant using equation (14).
7. Update pheromone using equation (16).
8. end for
9. Update best solution
10. end for
11. Output the optimal route from the source node to the BS.

5.3. Cluster maintenance

In this study, the cluster maintenance is considered as one of the important phases to balance the
load between the clusters. The clusters nearer to the BS drains their energy too fast, due to the inter
cluster traffic. Therefore, the cluster maintenance phase is required to eliminate the node failure. This
leads to increase in the lifetime while transmitting the data from the source node to the BS. If the residual
energy of the CH crosses beyond the threshold level, the BOA algorithm is again initialized to cluster the
network. Then the CHs are selected from the clustering algorithm and then the ACO is used to obtain the
routing path through the CHs to BS.

In this proposed methodology, an effective CH selection is performed by BOA algorithm. The

CHs are selected by considering five different parameters node degree, node centrality, distance to its
neighbors, distance to the BS and residual energy. These parameters are used to select the optimal CH
from the set of nodes. The residual energy of the nodes is frequently monitored by BS to avoid the node
failure during data transmission. Subsequently, the optimal transmission path from the source node to BS
via CH that is obtained by using the ACO algorithm. It detects the shortest path to reduce the energy
utilization of the nodes. This BOA and ACO based optimal CH selection and route generation lead to
design energy efficient WSN. Therefore, an energy efficient WSN is used to improve the network lifetime
and increases the total number of packets transmitted to the BS during data transmission.

6. Performance evaluation

The experimental setup and performance of the proposed method are discussed in this section.
The performance of the cluster based routing is analyzed using various metrics alive nodes, dead nodes,
average energy consumption, delay, total packet sent, throughput, FND, HND and LND. The
performance measures are explained as follows:

FND: FND defines the number of rounds when the first node of the network dies. This FND specifies the
number of time for which all the nodes are in a fully operative manner.

HND: HND defines the number of rounds when 50% of nodes are dead in the network. The network
performs the data transmission even after half of the nodes are dead in the network.
LND: LND defines the number of rounds when all nodes of the network dies. This LND specifies the
number of rounds when the network becomes inoperative.

Alive nodes: Alive nodes define the number of nodes that are alive in the network. The network
performance is improved, when the network has a high number of alive nodes.

Average energy consumption: It defines the average amount of energy consumed by each node during
each iteration.

Total packets transmitted to the BS: The total packets transmitted to BS are directly proportional to the
alive nodes and residual energy of the nodes. The total packets received by BS are high when alive nodes
are high.

Throughput: The throughput is defined as the amount of bits transmitted to BS over WSN. Throughput
is measured as bits per second.

Packet drop ratio: The packet drop ratio is defined as the volume of packet loss occurred during the
transmission from source node to the BS.

Routing overhead: The routing overhead is defined as the ratio between total amounts of received
packets by BS to the total amount of generated packets.

6.1. Simulation setup

The proposed energy efficient routing protocol is implemented and validated in MATLAB
R2018a which is operated in a Windows 8 operating system with an Intel core i3 processor and 4GB
RAM. The main reason for using MATLAB is that the ease of mathematical operations and appropriate
data analysis. There are 300 sensor nodes randomly deployed in the sensing area of 200m×200m. The
first order radio model is considered as energy model to analyze the proposed method with other routing
protocols. The parameters of first order model are and
. The simulation parameters considered for the experimentation are presented in the
following Table 1. The objective of this study is to reduce the overall energy consumption of each node in
the network. So, the cluster based routing is developed with BOA based cluster head selection and ACO
based routing between the cluster heads. The inputs given to BOA for better CH selection are residual
energy, node degree, distance to the neighbors, distance to the BS and node centrality. In addition, the
inputs to the ACO are residual energy, distance and node degree. Moreover, the parameters of the BOA
and ACO are shown in the Table 2 and Table 3 respectively. The amount of time taken for CH selection
and cluster creation process are 10.4841s and 0.037804s. This proposed method is compared with some
traditional approaches like LEACH and DEEC, because these approaches are generally used for
improving energy efficiency of the WSN. Additionally, the proposed method is compared with some
existing methodologies FUCHAR [45], CRHS [34], BERA [36], CPSO [38], ALOC [39] and FLION
[42].
 Table 1. Simulation parameters

Parameters Value
Area 200m×200m
E 0.5J

Packet size 4000 bits

Table 2. Parameters of BOA

Parameters Value
Population size 100
Switch probability 0.8
Power exponent 0.1
Sensory modality 0.01
 Table 3. Parameters of ACO

Parameters Value
Number of ants 100
Pheromone exponential weight 1
Heuristic exponential weight 1
Evaporation rate 0.1

Initially, the comparison of the proposed methodology with existing algorithms is presented. The
existing algorithms considered to evaluate the proposed methodology are LEACH and DEEC. LEACH is
generally a popular and standard clustering protocol. This evaluation is accomplished in two different
scenarios that is related to the location of the base station. These two scenarios are specified as follows:

Scenario 1: In the first scenario, the BS is positioned at the center of the area (100,100) which
used to analyze the short range communications.

Scenario 2: In the second scenario, the BS is positioned out of the area (100,250). The position of
the BS (100,250) shows that it is located far from the sensing area and this scenario 2 is used to analyze
the long range communications.

The illustration of scenario 1 and scenario 2 is shown in the Figure 3 and Figure 4 respectively. In
addition, the performance of the proposed methodology is evaluated with different number of nodes in
scenario 1. The nodes variation considered in this simulation are 100 and 150.
Figure 3. Illustration of Scenario 1
Figure 4. Illustration of Scenario 2
 6.1.1. Performance evaluation in terms of alive nodes

The alive nodes performance of the proposed methodology is tested for different scenarios and for
different number of nodes with existing algorithms like LEACH and DEEC. For instance, the proposed
methodology and existing algorithms are simulated for 300 sensor nodes. The performance evaluation of
alive nodes for the proposed methodology with LEACH and DEEC algorithms is illustrated in figure 5
and figure 6, these shows the alive nodes performance for scenario 1 and scenario 2 respectively. Figure 5
shows that the number of alive nodes of the proposed methodology is better compared to the LEACH and
DEEC, when the BS is located at the center of the sensing area. Figure 6 shows better alive node
performance compared to the LEACH and DEEC algorithms, even when BS is located far from the
sensing area (100, 250). The proposed methodology prolongs the alive nodes for higher amount of rounds
based on its cluster maintenance and optimal routing through the WSN. The cluster maintenance phase of
the proposed methodology is used to maintain the network without any dead nodes until 6394 rounds.
Subsequently, the node with higher residual energy is taken into consideration for selecting the CH from
the group of sensor nodes. Hence, the alive nodes of the proposed methodology sustain until 8563 rounds
for scenario 1, but the LEACH and DEEC are sustained until 1397 and 1398 respectively. The figure 7
and figure 8 shows a comparison of alive nodes for the network with 100 nodes and 150 nodes
respectively. The figure 7 and figure 8 shows that the alive nodes of the proposed methodology are high
when compared with the LEACH and DEEC. The reason for lesser alive nodes of the LEACH is the
single hop data transmission that leads to consume higher energy through the WSN. In DEEC, the
probability of cluster head selection depends only on the amount of residual energy of the nodes.
Therefore, this alive node analysis presents that the proposed methodology has improved alive node‟s
performance compared to the LEACH and DEEC. The sensor node‟s balanced energy consumption of the
proposed methodology is the main reason for higher number of alive nodes. The balanced energy
consumption of the nodes is achieved by selecting the shortest path between the source node and the BS.
Figure 5. Performance evaluation of alive nodes for Scenario 1
Figure 6. Performance evaluation of alive nodes for Scenario 2
 Figure 7. Performance evaluation of alive nodes for 100 nodes

Figure 8. Performance evaluation of alive nodes for 150 nodes

6.1.2. Performance evaluation of average energy consumption

This section presents the performance comparison of the average energy consumption by the
proposed methodology with existing algorithms. The evaluation of average energy consumption by
proposed methodology with LEACH and DEEC is presented in figure 9, figure 10, figure 11 and figure
12. Figure 9 and Figure 10 present the average energy consumption for BS located at the center (100,
100) and far from the sensing area (100, 250) respectively. Figures 11 and Figure 12 present average
energy consumption of the network with 100 nodes and 150 nodes respectively. From these figures, it is
concluded that the proposed methodology achieves higher energy efficiency compared to the other two
algorithms LEACH and DEEC. LEACH consumes higher energy because of single hop data transmission
and random CH selection. The DEEC consumes higher energy because it does not consider the distance
during CH selection. The reason of the higher energy efficiency of the proposed methodology is an
optimal CH selection between the group of nodes and optimal path generation from CH to BS. In this
study, energy consumption of the nodes is reduced by selection of the next hop node with lesser distance.
Because, the energy consumption of node is directly proportional to the transmission distance between the
nodes.

Figure 9. Performance evaluation of average energy consumption for Scenario 1

Figure 10. Performance evaluation of average energy consumption for Scenario 2
Figure 11. Performance evaluation of average energy consumption for 100 nodes

Figure 12. Performance evaluation of average energy consumption for 150 nodes
 6.1.3. Performance evaluation of total packets transmitted to the BS

The total packets transmitted to BS are evaluated for two different scenarios and varying nodes.
The performance evaluation of packets transmitted to the BS of the proposed methodology, LEACH and
DEEC algorithms is presented in figure 13 and figure 14. The proposed methodology achieves a higher
number of received packets at the BS compared to the LEACH and DEEC for both scenario 1 and
scenario 2. The reason of the achievement of higher data packets at BS is the efficient fitness functions of
BOA and ACO (i.e., described in the section 5.1.3 and 5.1.5) of the proposed methodology. These fitness
functions are considered preserving the residual energy of the nodes leads to maximizing the number of
alive nodes through the WSN which leads to increase in the number of data packets transmitted to the BS.
Moreover, the fitness function of the BOA and ACO is used to select only the node with enough energy
to transmit the data packets to the BS. This helps to avoid the packet drop while transmitting the data
packets to the BS.
Figure 13. Performance evaluation of total packets transmitted to the BS for scenario 1 and scenario 2
Figure 14. Performance evaluation of total packets transmitted to the BS for different number of nodes
 6.1.4. Performance evaluation of throughput

This section presents comparison of throughput using the proposed methodology and existing
algorithms. The evaluation of throughput for proposed methodology with LEACH and DEEC is depicted
in figure 15 and figure 16. This evaluation is analyzed for two different scenarios those are placement of
BS at center (Scenario 1) and placement of BS far from the field (Scenario 2). Figure 15 and figure 16
illustrates that the proposed methodology has a higher throughput compared to the other two algorithms.
The reason behind lesser throughput is the sensors of the LEACH and DEEC, which lost their energy
faster than the proposed methodology. The inappropriate CH selection of LEACH and DEEC is the main
reason of higher energy consumption. The proposed methodology transmits higher data bits compared to
the existing algorithms due to the energy efficient CH selection and optimal route generation. The
proposed methodology with high throughput is achieved by successfully transmitting the data packets to
the BS.

Figure 15. Performance evaluation of throughput for scenario 1 and scenario 2

 Figure 16. Performance evaluation of throughput for different number of nodes

6.1.5. Performance evaluation of network lifetime

This section presents comparison of the network lifetime between the proposed methodology and
existing algorithms. In this study, the lifetime of a network is defined as the number of rounds when the
total number of nodes completely exhausts their energy. Three different metrics FND, HND and LND are
analyzed. The FND, HND and LND of scenario 1 and scenario 2 for the evaluation of the proposed
methodology with LEACH and DEEC is presented in Figure 17, 18 and 19 respectively. Moreover,
comparison of lifetime in terms of LND for different nodes is depicted in the figure 20. It is known that
the death of the first node doesn‟t affect the performance of the network, but the network‟s performance
become worse when 50% nodes dead during the data transmission. Moreover, the operation of the
network fully stop when the last node dies in the network. Evaluation of the network lifetime is performed
for scenario 1 and scenario 2. Figure 17 to figure 20 illustrate that the proposed methodology has
improved network lifetime due to its energy efficient CH selection and optimal path generation. An
optimal CH selection and path generation preserves the energy of the sensor nodes in the WSN which
used to improve the network lifetime during data transmission.
Figure 17. First node dies for scenario 1 and scenario 2

Figure 18. Half node dies for scenario 1 and scenario 2

Figure 19. Last node dies for scenario 1 and scenario 2

Figure 20. Last node dies for different number of nodes

 6.1.6. Performance evaluation of packet drop ratio and routing overhead

This section discusses the comparison of the packet drop ratio and routing overhead between
proposed method and existing algorithms. The evaluation of the packet drop ratio for different number of
nodes is presented in figure 21. Figure 22 shows the evaluation of routing overhead for different number
of nodes. Figure 21 and figures 22 show that the proposed methodology achieves less packet drop ratio
and routing overhead compared to the existing algorithms. The fitness function considered in the
proposed methodology helps to minimize losses occurred during the data transmission. The data packets
received by the BS are increased by avoiding the node failure in path generation. Moreover, the LEACH
and DEEC causes high packet drop ratio and routing overhead due to an inappropriate CH selection and
lesser network lifetime.

Figure 21. Performance evaluation of packet drop ratio for different number of nodes
 Figure 22. Performance evaluation of routing overhead for different number of nodes

6.2. Comparative analysis of the proposed methodology with other cluster based routing
algorithms

The performance of the proposed methodology is compared with some existing cluster based routing
algorithms to find out the effectiveness of the proposed methodology. The existing algorithms considered
in evaluating the proposed methodology are FUCHAR [45], CRHS [34], BERA [36], CPSO [38], ALOC
[39] and FLION [42]. The comparative results are also taken for different scenarios that are specified as
follows:

Scenario 3: In scenario 3, 200 sensors are deployed in the sensing area of 200×200m2. Here, the BS is
located at (250, 250). The scenario 3 is performed to compare the proposed methodology with CRHS [34]
and BERA [36].

Scenario 4: In scenario 4, 100 sensors are deployed in the sensing area of 100×100m2. In this case, BS is
located at (175, 50). The scenario 4 is performed to compare the proposed methodology with CPSO [38].

Scenario 5: In scenario 5, 100 sensors are deployed in the sensing area of 100×100m2. In this case, the
BS is located at (0, 0). This scenario 5 is performed to compare the proposed methodology with ALCO
[39].

Scenario 6: In scenario 6, 100 sensors are deployed in the sensing area of 100×100m2. In this case, the
BS is located at (50, 50). This scenario 4 is performed to compare the proposed methodology with FLION
[42].
 Table 4. Comparative analysis of residual energy for proposed methodology with existing routing
algorithms
Residual energy (J)
Rounds Scenario 2 Scenario 3 Scenario 4 Scenario 5
FUCHAR CRHS BERA Proposed CPSO ALOC Proposed
[45] [34] [36] [38] [39]
200 - 230.1352 350 388.1296 - 46 46.1878
400 - 140.3814 290 377.2473 - 43 42.5509
600 - 89.2002 250 365.7348 - 38 38.7841
800 - 57.4633 200 354.5350 - 35 35.0420
1000 - 34.6276 165 343.7332 - 30 31.3138
1200 - 24.6593 140 332.4183 - 27 27.5431
1400 - 18.5922 110 320.1775 - 23 23.8564
1500 - 15.9089 105 314.7801 - 20 22.0474

Table 5. Comparative analysis of alive nodes for proposed methodology with existing routing algorithms
Round Scenario 2 Scenario 3 Scenario 4 Scenario 5
s FUCHA Propose CRH BER Propose CPS Propose ALO Propose
R [45] d S [34] A [36] d O d C [39] d
[38]
200 300 300 129 160 200 100 100 100 100
400 300 300 77 128 200 100 100 100 100
600 300 300 43 114 200 100 100 100 100
800 300 300 30 74 200 0 100 100 100
1000 300 300 17 52 200 0 100 100 100
1200 300 300 12 40 200 0 100 100 100
1400 190 300 9 30 200 0 100 100 100
1500 70 300 7 26 200 0 100 100 100
2000 0 300 - - 200 0 100 100 100
2500 0 200 - - 152 0 99 0 63
Table 6. Comparative analysis of packets transmitted to the BS for proposed methodology with existing
routing algorithms
Packets transmitted to the BS (Byte)
Packets transmitted to the BS Packets transmitted to the BS
before HND
Scenario 2 Scenario 3 Scenario 4
FUCHAR [45] Proposed CRHS [34] BERA Proposed CPSO Proposed
[36] [38]
5928

Table 7. Comparative analysis of proposed methodology with FLION for scenario 6

Rounds Alive nodes Average energy consumption (J)
FLION [42] Proposed FLION [42] Proposed
500 100 100 0.35 0.48
1000 80 100 0.15 0.41
1500 48 100 0.05 0.34
2000 38 100 0.03 0.28
 Table 4, Table 5 and Table 6 present the comparative analysis of the proposed methodology with
other routing protocols in terms of residual energy, alive nodes and packets transmitted to the BS
respectively. Moreover, the Table 7 shows the comparative analysis of the proposed methodology with
FLION [42] in terms of alive nodes and average energy consumption. The residual energy of the
proposed methodology is high when compared with the CRHS [34], BERA [36] and ALOC [39] that is
shown in Table 4. It shows that the proposed methodology achieved higher energy efficiency compared to
other routing algorithms. The reason of better performance of the proposed methodology is optimal CH
selection and route generation during the data transmission. Table 5 shows that the proposed methodology
has higher number of alive nodes compared to other routing algorithms. For example, entire nodes in the
network with FUCHAR [45] are dead before round 2000, but in the proposed methodology the last node
is dead at the round of 2753. The proposed methodology achieves a higher number of received data
packets in BS which is presented in Table 6. For example, the proposed method achieves of
data packets in BS that is high when compared with the CRHS [34] and BERA [36]. The received packets
of CRHS [34] and BERA [36] are and respectively. Table 7 shows the proposed
methodology has improved performance than the FLION [42] algorithm. For example, the alive nodes of
the proposed method is 100 at 2000 rounds and it is high when compared to the FLION [42] algorithm. In
scenario 6, the FND of the proposed methodology is 2519th round which shows that the lifetime of the
proposed methodology is improved compared to the FLION [42] algorithm. The fitness functions used in
the proposed methodology are the reason to achieve higher number of data packets in BS. The residual
energy of the nodes is preserved based on this fitness function and it leads to increase in number of alive
nodes over WSN. Thus, a large number of alive nodes results in higher amount of data packets at BS. In
addition, the throughput of the proposed methodology for scenario 5 is is high when
compared with ALCO [39]. The proposed methodology transmits high amounts of bits due to its high
lifetime compared to the other routing algorithms.

7. Conclusion

The optimal CH selection and route generation are considered as challenging tasks in WSN. In this
study, a combination of BOA and ACO was used to reduce the overall energy consumption to increase
the lifetime of the network. BOA based CH selection was performed using five different parameters
residual energy of the nodes, distance to the neighbors, distance to the Base Station (BS), node degree and
node centrality. This fitness function was used to select the optimal CH through the group of nodes. The
energy efficient route selection was obtained by optimizing ACO with three different parameters distance,
residual energy and node degree. During the simulation of the proposed methodology, the BS moved
from inside to outside to know its performance. The comparison between proposed methodology and
existing algorithms shows that the proposed methodology had a higher network lifetime compared to
LEACH and DEEC. The proposed methodology was also compared with existing routing approaches
such as FUCHAR, CRHS, BERA, CPSO, ALOC and FLION. It showed that the proposed methodology
had higher network performances compared to the existing algorithms. For example, the alive nodes of
the proposed methodology were 99 at 2500 rounds that was high when compared with CPSO method.
Furthermore, WSN performances can be improved by using novel optimization algorithms.
Declaring

Funding: This study was not funded by any organization.

References

[1] Vajdi, A., Zhang, G., Wang, Y. and Wang, T., 2016. A new self-management model for large-scale
event-driven wireless sensor networks. IEEE Sensors Journal, 16(20), pp.7537-7544.

[2] Kaushik, A., Indu, S. and Gupta, D., 2019. A Grey Wolf Optimization Approach for Improving the
Performance of Wireless Sensor Networks. Wireless Personal Communications, 106(3), pp.1429-1449.

[3] Kevin, P., & Samarakoon, U. T. (2019). Performance analysis of wireless sensor network localization
algorithms. International Journal of Computer Networks and Applications (IJCNA), 6(6), 92-99.

[4] Sabet, M. and Naji, H.R., 2015. A decentralized energy efficient hierarchical cluster-based routing
algorithm for wireless sensor networks. AEU-International Journal of Electronics and Communications,
69(5), pp.790-799.

[5] Mann, P.S. and Singh, S., 2019. Improved metaheuristic-based energy-efficient clustering protocol
with optimal base station location in wireless sensor networks. Soft Computing, 23(3), pp.1021-1037.

[6] Prasad, D.R., Naganjaneyulu, P.V. and Prasad, K.S., 2017. A hybrid swarm optimization for energy
efficient clustering in multi-hop wireless sensor network. Wireless Personal Communications, 94(4),
pp.2459-2471.

[7] Usha, M., J. Sathiamoorthy, R. Ashween, and Bhagavath Nishanth Ramakrishnan. "EEMCCP-A
Novel Architecture Protocol Design for Efficient Data Transmission in Underwater Acoustic Wireless
Sensor Network." International Journal of Computer Networks and Applications, 7(2), PP:28 – 42, 2020.

[8] RejinaParvin, J. and Vasanthanayaki, C., 2015. Particle swarm optimization-based clustering by
preventing residual nodes in wireless sensor networks. IEEE sensors journal, 15(8), pp.4264-4274.

[9] Osei, A. I., & Isaac, A. M. Partial Topology-Aware Data Distribution within Large Unmanned Surface
Vehicle Teams. International Journal of Computer Networks and Applications, 7(2), PP:19 – 27, 2020.

[10] Wang, L., Qi, J., Xie, W., Liu, Z. and Jia, Z., 2017. An enhanced energy optimization routing
protocol using double cluster heads for wireless sensor network. Cluster Computing, pp.1-12.

[11] Haile, G., Yohannes, D. and Ababu, K., Lane Id based Selective Emergency Message Forwarding
Scheme for VANET.

[12] Jha, S.K. and Eyong, E.M., 2018. An energy optimization in wireless sensor networks by using
genetic algorithm. Telecommunication Systems, 67(1), pp.113-121.

[13] Mittal, N., 2019. Moth flame optimization based energy efficient stable clustered routing approach
for wireless sensor networks. Wireless Personal Communications, 104(2), pp.677-694.

[14] Rao, P.S., Jana, P.K. and Banka, H., 2017. A particle swarm optimization based energy efficient
cluster head selection algorithm for wireless sensor networks. Wireless networks, 23(7), pp.2005-2020.

[15] Wang, T., Zhang, G., Yang, X. and Vajdi, A., 2018. Genetic algorithm for energy-efficient clustering
and routing in wireless sensor networks. Journal of Systems and Software, 146, pp.196-214.
[16] Mohamed, K.S., Nouh, S.A.E.H. and Naguib, A.M., Adaptive Power Transmission and Efficient
Energy Scheme of DSR Protocol (APEE-DSR).

[17] J.Sathiamoorthy, Dr.B. Ramakrishnan “A competent three-tier fuzzy cluster algorithm for enhanced
data transmission in cluster EAACK MANETs “Springer Soft Computing, Volume 22, pages 6545–6565
(2018).

[18] Mohajerani, A. and Gharavian, D., 2016. An ant colony optimization based routing algorithm for
extending network lifetime in wireless sensor networks. Wireless Networks, 22(8), pp.2637-2647.

[19] Mittal, N., Singh, U. and Sohi, B.S., 2019. An energy-aware cluster-based stable protocol for
wireless sensor networks. Neural Computing and Applications, 31(11), pp.7269-7286.

[20] Moh‟d Alia, O., 2018. A dynamic harmony search-based fuzzy clustering protocol for energy-
efficient wireless sensor networks. Annals of Telecommunications, 73(5-6), pp.353-365.

[21]Yuste-Delgado, A.J., Cuevas-Martinez, J.C. and Triviño-Cabrera, A., 2019. EUDFC-Enhanced

Unequal Distributed Type-2 Fuzzy Clustering Algorithm. IEEE Sensors Journal, 19(12), pp.4705-4716.

[22] J.Sathiamoorthy, Dr.B. Ramakrishnan & M. Usha “A Three Layered Peer-to-Peer Energy Efficient
Protocol for Reliable and Secure Data Transmission in EAACK MANETs“Springer Journal of Wireless
Personal Communication, September 2018, Volume 102, Issue 1, pp 201–227.

[23] J.Sathiamoorthy, Dr.B. Ramakrishnan & M. Usha “A Trusted Waterfall Framework Based Peer to
Peer Protocol for Reliable and Energy Efficient Data Transmission in MANETs “Springer Journal of
Wireless Personal Communication, September 2018, Volume 102, Issue 1, pp 95–124.

[24] Mann, P.S. and Singh, S., 2017. Improved metaheuristic based energy-efficient clustering protocol
for wireless sensor networks. Engineering Applications of Artificial Intelligence, 57, pp.142-152.

[25] Singh, R. and Verma, A.K., 2017. Energy efficient cross layer based adaptive threshold routing
protocol for WSN. AEU-International Journal of Electronics and Communications, 72, pp.166-173.

[26] Kaur, S. and Mahajan, R., 2018. Hybrid meta-heuristic optimization based energy efficient protocol
for wireless sensor networks. Egyptian Informatics Journal, 19(3), pp.145-150.

[27] Khabiri, M. and Ghaffari, A., 2018. Energy-aware clustering-based routing in wireless sensor
networks using cuckoo optimization algorithm. Wireless Personal Communications, 98(3), pp.2473-2495.

[28] Wang, Z.X., Zhang, M., Gao, X., Wang, W. and Li, X., 2017. A clustering WSN routing protocol
based on node energy and multipath. Cluster Computing, pp.1-13.

[29] Haseeb, K., Bakar, K.A., Ahmed, A., Darwish, T. and Ahmed, I., 2017. WECRR: Weighted energy-
efficient clustering with robust routing for wireless sensor networks. Wireless Personal Communications,
97(1), pp.695-721.

[30] Haseeb, K., Bakar, K.A., Abdullah, A.H. and Darwish, T., 2017. Adaptive energy aware cluster-
based routing protocol for wireless sensor networks. Wireless Networks, 23(6), pp.1953-1966.
[31] Haseeb, K., Bakar, K.A., Abdullah, A.H., Ahmed, A., Darwish, T. and Ullah, F., 2016. A dynamic
Energy-aware fault tolerant routing protocol for wireless sensor networks. Computers & Electrical
Engineering, 56, pp.557-575.

[32] Han, G. and Zhang, L., 2018. WPO-EECRP: Energy-Efficient Clustering Routing Protocol Based on
Weighting and Parameter Optimization in WSN. Wireless Personal Communications, 98(1), pp.1171-
1205.

[33] Morsy, N.A., AbdelHay, E.H. and Kishk, S.S., 2018. Proposed Energy Efficient Algorithm for
Clustering and Routing in WSN. Wireless Personal Communications, 103(3), pp.2575-2598.

[34] Lalwani, P., Das, S., Banka, H. and Kumar, C., 2018. CRHS: clustering and routing in wireless
sensor networks using harmony search algorithm. Neural Computing and Applications, 30(2), pp.639-
659.

[35] Lalwani, P., Banka, H. and Kumar, C., 2017. CRWO: Clustering and routing in wireless sensor
networks using optics inspired optimization. Peer-to-Peer Networking and Applications, 10(3), pp.453-
471.

[36] Lalwani, P., Banka, H. and Kumar, C., 2018. BERA: a biogeography-based energy saving routing
architecture for wireless sensor networks. Soft Computing, 22(5), pp.1651-1667.

[37] Rao, P.S. and Banka, H., 2017. Novel chemical reaction optimization based unequal clustering and
routing algorithms for wireless sensor networks. Wireless Networks, 23(3), pp.759-778.

[38] Mekonnen, M.T. and Rao, K.N., 2017. Cluster optimization based on metaheuristic algorithms in
wireless sensor networks. Wireless Personal Communications, 97(2), pp.2633-2647.

[39] Yogarajan, G. and Revathi, T., 2018. Improved cluster based data gathering using ant lion
optimization in wireless sensor networks. Wireless Personal Communications, 98(3), pp.2711-2731.

[40] Ezhilarasi, M. and Krishnaveni, V., 2019. An evolutionary multipath energy-efficient routing
protocol (EMEER) for network lifetime enhancement in wireless sensor networks. Soft Computing,
23(18), pp.8367-8377.

[41] Gao, F., Luo, W. and Ma, X., 2019. Energy constrained clustering routing method based on particle
swarm optimization. Cluster Computing, 22(3), pp.7629-7635.

[42] Sirdeshpande, N. and Udupi, V., 2017. Fractional lion optimization for cluster head-based routing
protocol in wireless sensor network. Journal of the Franklin Institute, 354(11), pp.4457-4480.

[43] Dahiya, S. and Singh, P.K., 2018. Optimized mobile sink based grid coverage-aware sensor
deployment and link quality based routing in wireless sensor networks. AEU-International Journal of
Electronics and Communications, 89, pp.191-196.

[44] Xie, W.X., Zhang, Q.Y., Sun, Z.M. and Zhang, F., 2015. A clustering routing protocol for WSN
based on type-2 fuzzy logic and ant colony optimization. Wireless Personal Communications, 84(2),
pp.1165-1196.
[45] Arjunan, S. and Sujatha, P., 2018. Lifetime maximization of wireless sensor network using fuzzy
based unequal clustering and ACO based routing hybrid protocol. Applied Intelligence, 48(8), pp.2229-
2246.

[46] Arora, S. and Singh, S., 2019. Butterfly optimization algorithm: a novel approach for global
optimization. Soft Computing, 23(3), pp.715-734.

[47] Dorigo, M. and Gambardella, L.M., 1997. Ant colony system: a cooperative learning approach to the
traveling salesman problem. IEEE Transactions on evolutionary computation, 1(1), pp.53-66.
Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests:
Mrs. Prachi Maheshwari (Corresponding Author)

Prachi Maheshwari is currently pursuing her Ph.D. in Computer Science and Engineering
Department from the National Institute of Technology, Delhi, India. She received her Bachelor's
in Technology in 2005 from Faculty of Engineering and Technology, Agra College, Agra, India;
Masters in Technology from National Institute of Technology, Hamirpur, Himachal Pradesh,
India in 2010. Her research interest lies in the domain of application of Game theory and Graph
Theory in solving various problems in WSN and designing energy-efficient techniques for
wireless sensor networks.

Dr. Ajay K. Sharma (Co-Author)

Dr. Ajay K. Sharma is working as Vice-Chancellor IKG Punjab Technical University, Jalandhar,
Punjab, India. He received his B.E. in Electronics and Electrical Communication Engineering
from Punjab University Chandigarh, India in 1986, M.S. in Electronics and Control from Birla
Institute of Technology (BITS), Pilani in the year 1994 and Ph.D. in Electronics Communication
and Computer Engineering in the year 1999. His Ph.D. thesis was on „„Studies on Broadband
Optical Communication Systems and Networks‟‟. After serving various organizations from 1986
to 1995, he has joined National Institute of Technology (Erstwhile Regional Engineering
College) Jalandhar as Assistant Professor in the Department of Electronics and Communication
Engineering in the year 1996. From November 2001, he has worked as Professor in the ECE
Department and thereafter he has worked as Professor in Computer Science and Engineering
from 2007 to 2013 in the same institute. His major areas of interest are broadband optical
wireless communication systems and networks, dispersion compensation, fiber nonlinearities,
optical soliton transmission, WDM systems and networks, radio-over-fiber (RoF) and wireless
sensor networks and computer communication.

Dr. Karan Verma (Co-Author)

Dr. Karan Verma is working as Assistant Professor in Computer Science and Engineering
Department, NIT Delhi. He received his B.E and M. Tech degrees in Information Technology
and Computer Science & Engineering, respectively, from the University of Rajasthan in 2008,
and Indian Institute of Technology, Roorkee in 2010. He received his Ph.D. degree from
Department of Computer and Information Science at the Universiti Teknologi PETRONAS,
Malaysia, and Post DoC (PDF) in Computer Science & Engineering from University of
Californian, Davis, USA. His research interests include information security, wireless sensor
networks and VANET.