Mobile Sink in Fuzzy Logic and Meta-heuristic Firefly Algorithm based Routing Scheme to Extend Network Lifetime of WSN (MSFLMFLA)

Abstract

Nowadays, wireless sensor network (WSN) improves people's lives by assisting them with a variety of applications. The major challenge in WSN is consumption of power as well as lifetime of the network. Clustering is the important method for saving energy in WSN because the separation of the sender and the receiver is related to transmission energy. In this paper we used fuzzy rules for clustering. Wireless sensor networks are susceptible to energy holes, in which sensors nearer to a static sink rapidly lose energy. To solve this problem and extend network life we used a mobile sink (MS). In this paper, a customized mobile sink node called the mobile data sender (MDS) has been tried to introduce for collecting data from the sensors by going to visit every node and then going to send this to the base station. This paper proposes nature inspired heuristic discrete firefly algorithm to optimal way accumulate information from sensor nodes in order to decrease the path travelled by the MDS while doing the tour. So, in this paper, we consider Mobile Sink in Fuzzy Logic and Meta-heuristic Firefly Algorithm based routing scheme to extend network lifetime of WSN (MSFLMFLA). Here we compare throughput and network lifetime with the static sink in FLMFLA and EHR-DC and LEACH protocol. The results of simulation shows that the MSFLMFLA increases the throughput, residual energy and also increases the data received by the sink.

Full text

Mobile sink in Fuzzy Logic and Meta-heuristic Firefly Algorithm
based Routing Scheme to Extend Network Lifetime of WSN
(MSFLMFLA)
Pranali Navghare1 Sudhakar Pandey2 Deepika Agrawal3
Dept. Of Information Technology, NIT, Raipur, Raipur, India
pranalinaoghare@gmail.com, spandey.it@nitrr.ac.in, dagrawal.it@nitrr.ac.in
Abstract Nowadays, wireless sensor network (WSN) improves people's lives by assisting them with a variety of
applications. The major challenge in WSN is consumption of power as well as lifetime of the network. Clustering is
the important method for saving energy in WSN because the separation of the sender and the receiver is related to
transmission energy. In this paper we used fuzzy rules for clustering. Wireless sensor networks are susceptible to
energy holes, in which sensors nearer to a static sink rapidly lose energy. To solve this problem and extend network
life we used a mobile sink (MS). In this paper, a customized mobile sink node called the mobile data sender (MDS)
has been tried to introduce for collecting data from the sensors by going to visit every node and then going to send
this to the base station. This paper proposes nature inspired heuristic discrete firefly algorithm to optimal way
accumulate information from sensor nodes in order to decrease the path travelled by the MDS while doing the tour.
So, in this paper, we consider Mobile Sink in Fuzzy Logic and Meta-heuristic Firefly Algorithm based routing scheme
to extend network lifetime of WSN (MSFLMFLA). Here we compare throughput and network lifetime with the static
sink in FLMFLA and EHR-DC and LEACH protocol. The results of simulation shows that the MSFLMFLA increases
the throughput, residual energy and also increases the data received by the sink.
Keywords WSN (wireless sensor Network), fuzzy rules, firefly algorithm, MDS (mobile data sender), MS (mobile
sink)
1 Introduction
Since the last decade's wireless sensor network applications start increasing in multiple domains because of its
advancement in technology. A wireless sensor network is a network of sensor nodes that are frequently embedded at
random in a specific area over dynamic environments. Such nodes can detect, process, and transmit data to surrounding
nodes and the sink node (SN). Furthermore, many such tiny sensor nodes have limited capacities including limited
memory, low data processing, low storage, and, another very importantly, a minimal power unit [1]. Figure 1
represents the architecture of sensor network [37]. Sensor nodes will mostly lose energy after receiving and
transmitting data most of the times because nodes can be deployed with initial energy only once and cannot be charged
up.The primary function of nodes is to accumulate information from various nodes and transmit it to a main unit
known as sink to be collected [2]. WSN has now been frequently used for the defense department security, agricultural
and industrial pollution management, biological healthcare facilities, remote surveillance of the high risk area and
several other essential areas, the eventual commercial application has bring tremendous ease to individuals [3]. In
wireless sensor networks (WSNs), the coverage problem is termed as a measure of how well a sensor nodes is
monitored the network field. Over the years, this subject has sparked a huge interest, and therefore, several coverage
strategies have been developed [5].

Fig 1: Architecture of sensor network [37]
The time from the commencement of the operation of network till the first sensor node loose its energy is known as
network lifespan [8]. In wireless sensor networks (WSNs), the difficult problem of data aggregation is critical for
lowering traffic and overhead of the network . The majority of information supplied by sensor nodes is repetitive, and
performing procedures on it often results in higher power consumption and a shorter network lifetime [30]. Individual
sensor node is capable for transmit direct information to base station, but this result in higher energy consumption,
which ultimately affects the network's lifetime. In terms of reducing the networks total energy consumption. Clusters
are formed by grouping together nodes with similar characteristics or nodes that are nearby [4]. Clustering is one of
the protocol which saves energy in WSN so its significantly minimize the consumption of energy while also extending
network life. The clustering protocol divides the network into several clusters, each of which has a cluster head (CH)
and various cluster members (CMs) as shown in figure 2 [37]. The CH is capable of collecting and processing the
information acquired by the CMs, and sending the grafted data to the sink node [3]. Clustering protocol investigations
has kept going, focusing on the cluster head selection technique, because CH selection has a direct impact on network
energy consumption. One of the very first conventional clustering approaches for sensing devices, is a distributed,
adaptive clustering protocol where only CHs are formed at regular intervals using a probabilistic formula in every
round, and CH will declare to all nodes, and nodes then choose CH received from the sensors signal strength [3]. One
of the strategies used for the optimal cluster head selection is the nave bayes classifier, which has shown to be
extremely promising in terms of increasing the lifetime of nodes over time slices [10]. LEACH Protocol is an example
of a hierarchical routing algorithm. It's self-adapting and self-organizing. LEACH protocol employs rounds as its unit
of measurement, each round consists of a cluster predefined stage and a stable stage. In order to reduce needless energy
consumption, the stable stage would have to be much lengthier than the predefined stage [6].
Fig 2: wireless sensor nodes with cluster [37]

LEACH has some fundamental flaws, such as the fact that each station is connect directly to every node and the node
which is at center, that’s decreases its usage for smaller networks. TDMA plans for LEACH allocates schedule open
positions to each hub regardless of whether the data has to transfer by hub or not. This results in unnecessary
postponement and overhead [11]. WSNs have been subjected to a plethora of clustering algorithms and strategies. To
form clusters, K-means clustering algorithm employs by the KM-LEACH. K-means clustering algorithms both
centralized and distributed. LEACH-CKM considers remote nodes when forming groups using the classification
method of K-means and transmits data using a routing protocol which transmit minimum energy [14].
Many attributes, like energy of the node, degree, and the energy condition for the neighboring nodes in the network,
all has considered at the same time. When developing the WSN dispersed clustering algorithm, As a result, the
appropriate CH is chosen in the equilibrium of multi-condition has a significant impact on the overall stabilization of
the networks which has cluster. A fuzzy logic system, on the other hand, can only give proper explaination to this
type of multifactor problem evaluation, such as CHs selection. In several other words, the fuzzy logic system can take
into account a variety of clustering factors when selecting CHs [15]. Many factors influence WSN clustering
algorithms,such as the amount of energy remaining in sensor nodes and their total distance from their BS other factors,
on the other hand, can be considered if the problem is thoroughly investigated. As a result, clustering is more suitable
if the clustering algorithm wants to take advantage of more energy-affecting factors. The fuzzy Inference System (FIS)
is a powerful tool for combining parameters in order to improve parameter performance [24]. Author in this paper [7]
present a centralized fuzzy-based clustering algorithm for WSN energy-efficient routing algorithms. Generally, fuzzy
logic is employed to resolve the uncertainty in WSNs. With the help of fuzzy logic, a system can be configured to run
more efficiently with partial data .
A path of routing from a source node to a destination node can be determined using a variety of optimization
approaches [31]. Several algorithms including particle swarm optimization, differential evolution, the bat algorithm,
the firefly algorithm, and the cuckoo search, has emerged in the last two decades, and they had great promise in solving
technically difficult optimization problems. Between these new algorithms, the firefly algorithm was shown to be very
efficient in dealing with multisensory, global optimization techniques [20]. The Firefly algorithm is a recent
phenomenon in designed to simulate evolutionary algorithms. It is easy, but adequate process parallelly with high
strength, which allows it to be used in a variety of areas like dispatch of economy and cryptography [22]. The firefly
algorithm is a method which corresponds to the inspired by biological algorithm and help to solve the objective
function. The flashing nature of fireflies inspired this algorithm. The fireflies which are bright attracted towards the
less-bright fireflies, that used to locate the searching area [28]. Information received from all the sensor nodes is
collected and it’s transmitted to the base station these sensor nodes have a limited amount of energy and memory.
Because of the increased data relaying under multi-hop communication, Sensor nodes in close proximity to the base
station rapidly exhaust their battery energy, reducing the WSN’s lifetime of the network. Therefore, that strategy
called as the emerged problem of energy hole. As a result, the use of clustering and mobile sinks can be regarded as
an important method for reducing energy consumption [34]. By adding mobility on a sink node, energy consumption
can be significantly lowered. As a result, while transmitting data, mobile data collection can result in lower energy
utilization among wholly sensor nodes [36].
The remaining paper is arranged as follows: The research stuff in compression techniques and data cluster analysis
using firefly algorithms in WSNs is outlined through Section 2. The System model is described in Section 3. Section
4 discusses the proposed work, and Sections 5 and 6 explain the fundamentals of fuzzy rules and the firefly algorithm.
The firefly algorithm for mobile sink is explained in Section 7. Section 8 discusses the performance evaluation, and
Section 9 concludes the paper.
2 Related work
For environmental monitoring, an easy and productive clustering technique termed energy efficient structured
clustering algorithm (EESCA) is suggested in this study [9]. Cluster heads (CHs) are chosen using distance of
communication as well as energy remaining. The new methodology is shown to be favorable for WSNs when
compared to the well-known previously algorithms low energy adaptive clustering hierarchy (LEACH) and scaled
energy efficient clustering hierarchy (SEECH). The research [12] suggests employing clustering based on particle
swarm optimization and routing based on harmony search in WSN to improve network lifetime. As a result, this study

is established for global ideal cluster heads choosing and gateway nodes to decrease the CH's utilization of energy
while sending aggregated information to the base station (BS). The results findings that the PSOHSA algorithm
outperforms other traditional protocols like EEABR and IHSBEER as well as that increases the network's life. The
"fuzzy based unequal clustering algorithm" is studied in this study [13] to extend the network's life. This approach
creates clusters that are uneven in size. This is done to keep the energy consumption in check. The fuzzy logic
technique is used to choose cluster heads. The separation with base station, remaining energy, and density are all input
variables. When compared to its competitors, it extends the network's lifespan. The author of this study [16] presented
a cluster-based SPIN (CBSPIN) routing algorithms, that is an extended type of the SPIN routing protocol that uses a
clustering methodology to construct the network. The MATLAB platform is utilized to simulate both the SPIN and
CBSPIN protocols, and the results reveal that CBSPIN performs better as compared with time as well as energy.
The author considers topology of network based on complex optimization network theory for resolve the energy saving
issue of WSN in this research [17]. Here an energy-efficient WSN model was proposed. The clustering properties of
a small world network are similar to those of the network's rules, but it also resembles random networks with short
average path lengths. It’s used to improve the overall energy efficiency of networks. The simulation results support
this model's energy efficiency and the renovating of the sink nodes to assure the WSN's regular performance. The
study [18] describes a method for increasing the life of a network in a WSN by employing genetic algorithm-based
clustering and particle swarm optimization-based routing. A genetic algorithm uses distance and energy characteristics
to identify the optimum cluster head (CH) that gathers information from the remaining nodes. Furthermore, the
optimization algorithm of particle swarm is based on the selection of optimal routing paths for all relay nodes sending
data to the base station (BS). This procedure demonstrates that the relay node facilitates and supports communication
between the sink and CH, resulting in increased energy efficiency. To boost network performance, a fitted firefly
heuristic, the firefly algorithm, is suggested in this paper [19]. Simulated results show that the suggested methodology
outperforms LEACH and energy well organized hierarchical clustering.
A novel congestion control method depends on optimal rate is developed in this study [21] in order to deliver energy
efficient transmissions. To reduce energy usage across the network, a congestion control method based on cluster
routing is introduced. The rate control technique reduces the end-to-end delay, extending the life of the network over
a longer simulation period. Author proposed a routing which uses the multipath approach for homogeneous WSNs in
this paper [23]. Grouping the nodes in the network, finding the pathways within the CHs, and continuing the pathways
are the three phases of the proposed technique. In forms of end-to-end delay, utilization of energy, losing rate of
packet, and network lifetime the simulation demonstrate that this multipath routing exceeds other routing approaches.
The authors of this study [25] offer a Distributed Energy aware Fuzzy Logic based routing algorithm (DEFL) that
targets utilization of energy as well as balancing energy at the same time. For the shortest path calculation, this
approach captures network state using relevant energy measurements and maps them to cost values. The lifetime of
network produced by DEFL outperforms the best as compared to all studied systems under varied load of traffic levels,
according to simulation data. Using a free space energy consumption model, it is shown that single-hop forwarding
consumes minimum energy [26]. When compared to other methods, results demonstrate that the algorithm efficiently
increases the life of the network while also achieving energy planning and balance of the energy. A HWSNs routing
protocol depends on the grey wolf optimizer (HMGWO) is studied in this study [27]. The process begins by explaining
various functions of fitness for different nodes energy, after which the values for fitness of the nodes are determined
and used as starting weights in GWO. When compared with the different algorithms, the HMGWO protocol enhances
the network lifetime. In this research [29], the author focuses on contemporary hierarchical routing methods that rely
on the LEACH protocol for improving the performance as well as lifespan of WSN. As a result, the Node Ranked-
LEACH technique is presented. Based on the node rank algorithm, this suggested protocol improves the total network
lifetime.
In this research [32], a fuzzy based balanced cost CH selection algorithm (FBECS) is suggested, that takes into account
the residual energy, distance with the sink, and density of nodes in the region as inputs to the FIS. In terms of improved
period of stability, longer life with balancing of load, and large data sending to BS, the simulation demonstrate that
the FBECS is above as compared with BCSA and LEACH. In this study [33], a novel method of clustering known as
Energy Centers Searching using Particle Swarm Optimization (ECPSO) is provided to skipping the holes of energy

as well as seeking centers of energy for selection of CHs. The authors did a number of simulations to show that the
EC-PSO beats other works in terms of life of network extension as well as ratio of energy usage.
This paper [35] proposes an event-aware hierarchical routing with differential compression (EHRDC) scheme. It
groups sensors and chooses the cluster head (CH) in each group to handle routing and compression. In normal times,
sensors transmit information to their CHs, which are compressed by the use of spatial correlation. When events appear,
sensors adaptively forward data to nearby CHs to raise the efficiency of compression. Through simulations, the authors
show that EHRDC is outperforming other methods in terms of life of the network and the amount of sensed data
collected by the base station. However, in [35] the node which energy is higher is assigning the cluster head that might
be quite away with the sink node.
The firefly algorithm has the fewest parameters, is capable of automatic subdivision, and can interact with
multimodality. This prompted the use of the firefly algorithm, a distinct optimization technique, for the proposed
mobile information collection challenge in WSN. MDS is used for planning the tour and for collecting the
information in wait time, limited WSNs. In this case, the BS and sensor nodes are considered to just be stationary,
and the sensors have been installed at random across the selected area. An MDS starts at the base station and proceeds
among all sensors by using the perfect option obtained from the proposed technique. That’s decreases the MDS's total
transport costs as well as the power consumption of each sensor node. In this paper Mobile Sink in Fuzzy Logic and
Meta-heuristic Firefly Algorithm based routing scheme to extend network lifetime of WSN (MSFLMFLA) modify
the cluster head selection and forwarding data so as to increase network lifetime and the MSFLMFLA is used for
finding the best way for the MDS to take across the network while collecting sensor information across all nodes along
the path.
3 System Model
Consider a sensing field in which sensors are uniformly distributed and form a connected WSN. All sensors seem to
be similar in terms of hardware and battery capacity, and generate the same type of sensing data, allowing them to be
grouped and compacted. The sensing field is already divided into grids. Utilizing various positioning strategies
because sensor locations can be obtained, each sensor knows which grid it belongs to. Events can occur at any time
and in any location. Sensors would start generating sensing data whose values differed greatly from those generated
during normal times after detecting events.
A sensor expends energy in the production, transmission, and reception of data. When a sensor ni generates a m-bit
packet of sensing data, it expends energy.
  
 
 
   (1)
Where viG, ciG and tiG denote the voltage, current and time required by ni for generating the packet respectively. When
sensor node ni sends the packet to a node nj , ni takes an amount of energy :
   
   
(2)

is the power for ni’s transmitter and 
is the power for ni’s amplifier for sending a bit and  is the distance
function.
Suppose be the power for ni’s receiver to get a bit. Then nj spends an amount of energy to take the packet:
    (3)
WSN sensors detect data and send this one to the sink. When the separation between both the sink and sensor node is
greater, the sensor must expend more energy to transmit data. As a result, we introduced mobility to allow the sink
node to move all over the network, meet the sensors, and begin receiving one‘s information. Because the mobile sink
pays close attention to each sensor node, the power consumption within the sensors is drastically decreased.

The challenge of choosing the optimal tour of an MDS that comes to visit every nodes present in the network for
information collection is referred to here. The following is a mathematical description of an MDS's data collection
tour as an integer equation.
    



  (4)
          (5)
       (6)
  

     (7)
  

      (8)
              (9)
If a path exists among both sensors i and 'j,' sij is close to unity. The sensors are labeled to signifiers ranging between 1
and sN for I = [1,2 sN-1,sN]. According to equations 7 and 8, the MDS must begin arriving at each sensor from a
particularly clear other sensor, but that each sensor must depart from an absolutely clear other sensor. According to
equation 9, the MDS must have one tour that covers all sensors. The variables ui and uj appear to be dummy variables
in this model, and the following assumptions are made.
(1) Each practicable MDS sample consists of a single closed series of detectors. (2) For each tour taken by an MDS
in an attempt to cover all sensors,and (3) Incident information across all sensor devices makes a equal contribution in
the WSN applicationt. (4) It is assumed that the separation between the two sensors along each side is assumed to be
equal. An MDS performs the necessary information collection trip to gather information out of a network system with
deployed nodes. At the end of the tour, before reporting to base station the MDS tours all the sensors in the
most efficient order possible and collects one‘s information.
4 Proposed work
To conserve energy, the sensor node in WSN form clusters. So, each cluster does have a coordinator, abbreviated
cluster head. The sensor node collects data from the targeted region and start relays that data to the base station. Each
sensor node could serve both as a node which sense the information and a cluster head. The head of the cluster
gather information from its members and routes it all to the base station. As a result, the emphasis is on because large
amount of power has been drained from the deployed sN during information transmission.
The system architecture proposed in this work aims to implement sensor nodes within the impact zone by monitoring
the surroundings. The following assumptions are made in this proposed work:
1.The network consists of homogeneous sensor nodes with an initial energy of 6480 Joules.
2. Throughout the network the deployed sensor nodes are dispersed at random.
3. Once deployed, the sensor nodes have become static, i.e. they are immovable.
4. Sink node is mobile and base station is also static
5. After deployment, the sensor is unsupervised and energy constricted, i.e., the battery pack is irreplaceable.
6. RSSI (received signal strength index) computes the sensor to base station separation.
7. The Base station has an endless supply of power.
8. Adjusting the transmission power relative to the spacing is possible.
9. A symmetric radio communications network exists.

10. Sensor node inability occurs as a result of energy depletion.
11. The network continuously generates reports.
5 Fuzzy Rule in MSFLMFLA
The proposed methodology is intended to use fuzzy rules to modify the selection of cluster head. The cluster head is
chosen based on three variables: the node's residual energy, its distance from the grid's center, and the cost of
communication with the base station. Fuzzy rules are based on iterations such as LEACH. The framework of these
rules in a single round is divided into 2 phases. The first is the topology construction phase, and the second is the
phase of forwarding.
5.1 Topology construction phase
Each sensor node distributed throughout the targeted region. That once sensor node is installed inside the specific
location, the suggested technique plays a role. Before gathering information from target region, clusters are formed
following the formation of every round. In every round, maximum p percent cluster heads are selected from the
implemented alive sensor node. Following deployment, the sink node sends the packet into the targeted region (LOC
SN). Such a packet includes important information such as the sink node location. The coordinates of the classified
regions and the sensor node time slot to avoid collision. To avoid collision, the sensor node now broadcasts a message
throughout the network in order to comply with the time slot offered by the sink node. Once all broadcasts have been
completed, the sensor node calculates local parameters such as residual energy, distance from the grid's center, and
communication cost with the sink node. Algorithm 1 describes the cluster formation algorithm in the suggested
technique.
Algorithm 1: Formation of cluster in MSFLMFLA
Begin
SN ← The network's sensor nodes
ID← Sensor node identifier
SNode(x). P← Po
SNode(x). Position ← member
CH_table_value ← 0
CH_total← 1
SNode(x).PR ← Allocate probability depending on the region
SNode(x). N← Total number of nodes in under communication range (CR)
SNode(i). DTBS← Base station separation from SNode(x)
While (CH_total<=PR%)
{
Evaluate the Eligibility Index at each SNode (x).
Compute Average [EI]
Calculate the Threshold Value (THV) for SNode(x)
Create a random number (Rand) for each SNode (x).
If Rand < SNode(x). THV and CH_total<PR%

then
SNode(x). Position ← head of the cluster
CH_total++
Add SNode(x) to CH_table_value
end If
}
Each CH communicates with each SNode by sending a CH Message (i.e., RE).
To form a cluster, SNode (x) connects to the nearest CH.
Terminate
5.2 Model with Fuzzy Bases
Throughout portion, we introduce the fuzzy based method besides cluster head selection, as well as a technique of
clustering derived from the adaptive fuzzy system for effective WSN clustering. In WSN, a variety of factors influence
CH selection. As a result, they must be properly combined in order to make the right choices FIS would be an effective
method for accomplishing this goal. This enables the combination of all input parameters in such a manner that their
performance in cluster head selection is reflected. Here uses fuzzy logic to maximize the benefits of selecting the
cluster head. It is necessary to investigate the factors that influence cluster head selection. We use appropriate scales
to evaluate all of those aspects, as well as construct an effective fuzzy model with an efficient integration of fuzzy sets
and suitable parameters.
Fuzzy logic approach is used in this proposed method to select the best candidate for the position of cluster head.
Fuzzy logic would be effective at simulating experience of human and behavior of making the decision. Figure 3
depicts the basic framework of the fuzzy system used in the proposed work.
.
Fig. 3 Building blocks of Fuzzy System
It consists of four construction blocks:
Fuzzifier: In fuzzy based implementations, the system inputs seem to be crisp sets that must be transformed into fuzzy
sets. A degree of membership is allocated to every fuzzy set. Thus, fuzzifier is used to convert a crisp set into an
appropriate linguistic value.

Fuzzy Rule Base: This is made up with IF-THEN regulations that the individual specifies. The fuzzy system's dynamic
behavior is defined by a rule base with if-then statements. The knowledge base is another name for the fuzzy rule
base.
Inference Engine: An inference engine of fuzzy to input data and rules of IF-THEN simulates the system of human
inference. The inference engine of fuzzy is critical in drawing and inferring conclusions out from rule base conditions.
Defuzzification: The defuzzification system transforms its fuzzy set produced by the inference engine it into an output
of value from which some conclusions are drawn. The defuzzifier computes the probability based on the centroid.
a. Fuzzification
As input for fuzzy logic, we used three variables. Table 2 shows the fuzzifier crisp input variable, as well as its
maximum and minimum values for evaluating the eligibility index. The cost of communication with the base station
is the Euclidean distance from each sensor node with the base station. Distance from the centre of the grid count of
the distance from sensor node to it centre of that grid under consideration for CH candidature.
Table 1: Input function of Fuzzifier
Input
Linguistic Variables
Remaining
energy of the
node
Low Medium High
Distance to GS
Close Medium Far
Communication
cost with BS
(NPE)
Sparse Medium Dense
The FIS receives these crisp values (discrete values). These three variables are used to arbitrate the values of the
membership function (MF) and the point of intersection of the input factor, as shown in figures 4 to 6.
Fig 4: Variable of Input: Distance-to-GS

Fig. 5 Variable of Input NPE
Fig. 6 Variable of Input Remaining Energy
Measurement and Normalization of Linguistic Variables
A strategy used for CH election has a significant impact on the WSN's long life that is affected by a variety of
attributes. These attributes are expressed as linguistic variables as in framework with fuzzy logic. The proposed fuzzy
controller incorporates three linguistic variables. They have an impact on the lifetime of network either explicitly or
implicitly through each of 3 components: energy consumption by cluster heads, overall power consumption by non-
cluster head nodes, or power utilization loads via sensor nodes. The Min-Max normalization technique (shown in
equation (10)), is used for comparative optimization of linguistic variable values. The calculation is done comparable
to a discursive universe of zero to 800 based on their locations. Ultimately, values of a given variable of the sensor
node have been dispersed from across discourse universe based on its relative locations. As a result, the given variable
values have been normalized because they must allocate a highest valued sensor with the highest value for becoming a
cluster head. The remaining sensor nodes are prioritized based on the frequency with which they occur in comparison
to the variable values, i.e., minimum and maximum.
( ) ( ) ( ) ( ) ( )
* Normalized var Value var Min var Max var Min var= − −
(10)

In above equation Value (var) is the assigned variable's value when considering the most recent sensor and Min (var)
as well as Max (var) have become the assigned variable's values i.e., minimum and maximum between all sensor
nodes, respectively. It should be noted that when calculating any linguistic variable for a specific sensor, any
neighboring sensor closer from before the cluster head should be exempted. This is due to the fact that if this network
cluster is chosen, they will never be becoming members of it.
b. Fuzzy Rule Base
For conditions in IF-THEN, membership values of fuzzified variables fed into the rule base. A value is obtained by
applying the operators of fuzzy AND as well as OR to the input data. The method of aggregation unions every one of
the outcomes as well as selects the value which is to be maximum from the fuzzy set which is aggregated after applying
the 27 rules that shown Table 2. We have used inference system of Mamdani that is widely used because of its several
advantages, to calculate its eligibility index using fuzzy logic.
Table 2: Fuzzy Rules designed for Fuzzy logic.
Rule
NPE
Rem_Ener
gy
DTGS
EI
I
Dense
Low
Close
Good
II
Dense
Medium
Close
Better
III
Dense
High
Close
Far Better
IV
Sparse
Close
Low
Far Better
V
Medium
Medium
Mediu
m
Good
VI
Medium
High
Close
Best
VII
Medium
Low
Close
Better
VIII
Medium
Low
Close
Far Better
IX
Sparse
High
Close
Very Best
X
Dense
Low
Mediu
m
Bad
XI
Sparse
Medium
Close
Best
XII
Dense
High
Mediu
m
Good
XIII
Sparse
Low
Mediu
m
Good
XIV
Dense
Medium
Mediu
m
Fair
XV
Medium
High
Mediu
m
Better
XVI
Medium
Low
Mediu
m
Fair

XVII
Sparse
Medium
Mediu
m
Better
XVIII
Sparse
High
Mediu
m
Far Better
XIX
Dense
Low
Far
Worst
XX
Dense
Medium
Far
Worst
XXI
Dense
High
Far
Bad
XXII
Medium
Low
Far
Worst
XXIII
Medium
Medium
Far
Bad
XXIV
Medium
High
Far
Fair
XXV
Sparse
Low
Far
Bad
XXVI
Sparse
Medium
Far
Fair
XXVI
I
Sparse
High
Far
Good
c. Defuzzification
Table 3 depicts the variables of fuzzy that have been used to generate the output of crisp value.
Table 3: Output and Linguistic Variables
Output
Linguistic Variables
Eligibility Index
Better, Far Better, Fair, Good, Worst, Worse, Bad, Best, Very
Best
The centre of area (C*) technique is used for defuzzification, as described in equation 11.

 (11)
Membership functions come in a variety of shapes including triangular, trapezoidal, sigmoid, and gaussian. The only
requirement for an MF is that it must be in the range of 0 to 1. We used gaussian MF for intermediate levels in this
research schemes because its widely known when defining fuzzy set theory due to this lightness as well as values of
non-zero in all points, and Membership Function is Trapezoidal for variables which are in boundary due to the simple
design as well as computation is faster. Other membership functions can be used, but we found that membership
function of trapezoidal as well as gaussian produced the best results equation (12) as well as (13), respectively, give
the membership functions of the trapezoidal and gaussian being used our fuzzy inference system.
 
 (12)
     
 
  (13)

The trapezoid's feet and shoulders are denoted by p, q, and r, s, respectively.
The defuzzifier translates the accepted input into a crisp set as well as calculates the Eligibility Index of every node.
Fig. 7 Output-Var: Eligibility Index
After calculating the eligibility index for all nodes, compute the threshold (THV) using the equation 14.
 
 
 (14)
THV is computed using the mean of all sensor node eligibility indexes and the probability allotted to each sensor node
based on their presence in the detection area. For indiscriminate CH selection, each sensor node in a wireless network
randomly selects a number. Whenever the number will be below the computed THV, that sensor is considered for
taking the job of CH. The job of cluster head has been critical to the network's energy efficiency, and this is
repositioned for each round to balance the load even amongst the number of sensor nodes. Following the cluster head
selection process during the topology construction stage, nodes inside this cluster head item broadcast one‘s choice
inside the coverage area (Ca). By sending (join cluster member) towards the nearest cluster head, the node that was
not accepted for the cluster head role joins several of the nearby clusters in this manner it’s formed the clusters. If a
few nodes remain after cluster formation, they will participate a cluster closest with that sensor node. After that cluster
head acknowledges them. When all of the Sensor nodes are bound to clusters, the topology setup phase is complete.
5.3 Forwarding phase
Following the successful formation of a cluster during the setup phase of topology, cluster head is now expected to
gather information by their members in accordance with the allotted slot for TDMA to every member for
communicating with the free collision. During cluster formation, cluster heads assign a slot for TDMA to each of their
members. After collecting information from each sensor node, the cluster head integrates the data in terms of reducing
repeated data and thus reduce communication costs. The sink node collects the grouped information from each cluster
head in accordance with the TDMA schedule in order to ensure collision free communication. Along those same lines,
the suggested methodology completes one round successfully.
6 Firefly algorithm
Finding the shortest path from every cluster head to the base station is the goal of firefly algorithm. Along that a novel
function of fitness that consists of the residual energy of the node, distance from its center of the grid, and
communication cost with the sink node.
6.1 Initialization of Firefly and Representation

The fireflies in the firefly algorithm each represent a successful solution. Every firefly represents an information
transmitting path from a cluster head to the base station in routing. Each firefly has the same dimension as the overall
amount of cluster heads in the WSN, plus another additional place for base station. In this case, the scale of the firefly
is equivalent to p + 1. In which p denoted the amount of Cluster heads in the WSN as well as 1 represents the amount
of sink node.
Let   be the xth firefly, here every place          
is the (cluster head node) next hop in the data relay to the sink node.
6.2 Fitness Function Derivation
Our goal is to search the best path from every cluster head to base station. For accomplishing this, a fitness function
is designed that takes into account different parameter such as the residual energy of the node, distance from its center
of the grid, as well as cost communication with the sink node.
Residual energy: In the proposed work, the attractiveness factor will be computed based on the distance as well as
remaining energy of the relay node. This will make sure that the relay node not only is closest to the base station but
also has enough energy to gather data from the cluster head. In transmitting data, the relay node collects and merge
the information before transmitting it to the sink nodes. As a result, a higher residual energy relay node is a better
option. As a result, our first sub goal in terms of residual energy is f1 maximized as shown in equation 15.
(15)
Distance from its center of the grid: Distance from the centre of the grid makes sure that the node is well connected to
all the cluster members in the grid. If a node has fewer members cluster head, it will use less energy in communicating
with its members as well as will be able to sustain for a more time period. As a consequence, a node with a limited
node degree is a better option. As a result, the second sub goal in terms of node degree is to maximize f2 as shown in
equation (16).
  


 (16)
Communication cost with the sink node: Communication cost will make sure that the selected cluster head expends
optimal power in collecting the information from the cluster members and forward it to the base station. When the
separation from the cluster head with the base station is as short as possible, it will consume less energy. As a result,
the second goal is to reduce the separation from the cluster head with the base station. It will extend the life of the
network. As a result, the third sub objective in terms of distance is f3 maximized as shown in equation 17.
  


 (17)
However, in the Firefly optimization, the fitness of the node is determined by the attractiveness factor, which is
determined by the distance between the nodes in equation 18 and 19.
  
 
 (18)
  
 (19)
6.3 Displacement and Mobility
For every iteration, the fireflies which has least brightness move closer to the fireflies which has more brightness, as
well as the positions of every firefly are modified. If the newly modified place does not satisfy the scope, i.e. Is -ve
or larger than p + 1, the algebraic calculation fails. In this scenario, place is substituted with a number which is random
from 1 to p + 1. The preceding methods are repeated in iterative manner till reached the highest number.
1
1p
CHx
x
fE
=
=

6.4 Routing Algorithm
Procedure for determining a relatively close path
1. Firefly Initialization     
2. Determine the intensity of every firefly.
3. While    /*using (18) */
end
4. N! = high. Iteration do
The mobility of a firefly is determined by the intensity of the light.
5.While
!F
xM=
6. While  
7. If then
8. for    to p
Shuffle firefly to  to /*using (16) */
Compute the updated position and adjust the intensity of light.
end for
end
end
end
9. Rated the latest fireflies as well as identify the better one.
10. Compute Relay node ()       (i.e., path) use high (fitness ()).
7 Firefly algorithm for mobile sink
The original firefly algorithm was created to solve consistent optimization issues which could not be used to solve
complex optimization issues. Despite the fact that discrete firefly algorithm has numerous application domains, it has
yet to be used to handle the information collection issue in networks. As a result, in this article presents a firefly
algorithm for mobile data sender (MDS) to handle the information collection travel issue in networks while
minimizing MDS travel length. Furthermore, it is assumed that all sensor network data is equally significant and
distinct. This algorithm will boost the effectiveness of many sensor network health monitoring applications.
7.1 Explanation Illustration
The explanation for such information collecting travel issue used in this method is an appropriate pattern where in the
MDS must reach every sensor node for gathering one’s information. Every firefly in the initialization signifies the
first alternative to an information gathering travel issue. Every other item for this alternative denotes a node Id, as well
as the index denotes the position of the MDS's journey. As an example, a simplistic network with 10 nodes is assumed.
Every node is recognized by an identification numbered from 1 to 10. Figure 8 depicts one example of a firefly as
well as its way to solve. The MDS begins from the base station and travels sensor node 7 initially, collecting the
information, and after those travels to sensor node 4, collecting the information, and after those moves to sensor node

10, and so on until it reaches the base station, in which it updates every one of the gathered information as from sensors
for more handling and regulation.
Fig 8: Firefly Representation
7.2 Initialization of the Population
Depend on the number of nodes to still be installed throughout the sensor network as well as the number of fireflies,
the original sequence is created at random to use the function which is random. The rand iterative function iterates the
starting from 1 to SN, where ‘SN’ denotes the number of sensors. One such random possible combination denotes a
firefly explanation, and many these quantities of random variations have been produced and assumed as the original
solution dependent on the number of fireflies. The following is the procedure for initialization of population:
Initial_Population= []
For x =1: ffn
Variable = rand (Number of sensors)
TPopulation= Integrate (Initial_Population, Variable)
End
Return TPopulation
7.3 Intensity of Light
Because it depicts the illumination of the firefly, the intensity of light is indeed an attractive feature in the basic firefly
algorithm. Even though we understand, as in proposed method, every explanation of firefly includes the iterative
sequence of ID of sensors to just be attended. A firefly's intensity (I) is proportional to the average travel path length
relative to just the firefly's explanation. The Fi firefly's intensity is calculated by equation (20) as the fitness function.
Because the goal is to reduce the length of the data collection tour, the firefly’s intensity is calculated as the reciprocal
of the average travel path length, as shown in equation (20). That allows a firefly’s explanation for a shorter travel
path length for having a more intensity of light than another firefly. The firefly’s intensity dubbed 'Fi' has been
computed using equation (20).



 (20)
In this case, ‘Ii' denotes the Firefly’s Intensity ‘Fi' under considering, while ‘SN' denotes the number of sensors installed
in WSN.
7.4 Distance
The distance between fireflies in a continuous optimization problem is able to computed directly by using Euclidean
distance as well as used for optimization. Furthermore, because the information collecting travel issue is a distinct
optimization technique, the Euclidean distance approach will not be used. In this case, the separation from firefly F1

to F2 is measured as the proportion of the number of various edges in between those fireflies as well as the number of
sensors installed in the WSN, as calculated in equation 21.
 
 (21)
Where dF1,F2 denotes the separation from firefly F1 to F2, NE denotes the average amount of various edges from firefly
F1 and F2, and SN denotes the amount of deployed sensors. In this case, Figure 9 depicts the computation of the
distance from firefly F1 to F2. It’s discovered that edges (6, 9), (5, 2), and (8, 1) present in Fj are not available in Fi.
As a result, NE = 3 and SN = 10, and the distance dF1, F2 = 3/10 = 0.3.
Fig 9: Calculation of distance between two fireflies
7.5 Updating Solution set
The light intensity of every firefly in the explanation set is calculated. The firefly which is less bright is drawn to and
then travel to a firefly which is brighter. The mobility of a low bright firefly Fi in the direction of the brighter firefly
Fj is performed with an edge-based mobility. For that approach, the edges that are not available in a firefly Fi are
calculated, and a unknown edge is chosen at random. The edges in Fi were then moved away from the chosen absent
edge, reducing the separation from the firefly F1 to F2. As a result, every other firefly would then make contribution
‘k’ innovative solutions, where 'k' makes reference to the explanation change index. Firefly will be chosen as the
current offspring for iteration, as well as the process repeats till the stopping condition is achieved. Ultimately, the
better global firefly acquired provides the best route also for MDS to tour most of the sensors and obtain one’s
information while travelling the shortest route.
8 Evaluation of Performance
8.1 Setup of Simulation
In the proposed method we have been used MATLAB R2018a on a CPU of Intel Core i3 with RAM of 4 GB to
implement our MSFLMFLA. Experiments were carried out with layout sizes ranging from 1000 m to 1000 m in both
dimensions. 800 sensor nodes and 300 rounds have been taken. The effect of a sensor node's transmission range is
also investigated in 150 m. This paper considered two distinct scenarios. The first scenario assumes that the sink node
is static and placed in the right corner (scenario-I), and that the cluster head collects their data and forwards it to the
sink node based on the answers that provided by FLMFLA. The second scenario (scenario-II) is based on the
assumption that the sink node is mobile. MSFLMFLA chooses a bundle of backbone nodes and determines the best
route to each of them in order to collect their data. The parameters of the network and parameters of the firefly
algorithm used in the simulation are listed in Tables 4 and 5. MSFLMFLA's performance is evaluated by running
some existing routing techniques such as EHRDC and LEACH.
MSFLMFLA performance is evaluated and validated using FND (First Node Dead-Network stability), LND (Last
Node Dead-Network Lifetime), throughput, total alive nodes in every round, total dead nodes in every round, network
throughput, and network remaining energy during each round.

TABLE4.Simulation Parameters
Parameter
Value
Packet Length(m)
20 bytes
Battery Capacity
6480 J
Communication Range
150 m
Area of the Sensing Field
1000m * 1000m
Time interval
1800 seconds
Coefficients in Eq. (1)

=1.5V,
=25mA,
=0.25ms
Coefficients in Eq. (2)
=50nJ/bit,
=100J/bit (per
m2)
Coefficients in Eq. (3)
=50nJ/bit
Table 5: FA parameters
Parameters
Value
Number of fireflies
800
Number of Generations
300

1

0.2
8.2 Alive nodes and Dead nodes per round
Figure 10 depicts alive nodes with 0 to 300 sensors and two events in the sensing field. Figure 11 depicts the dead
counts of all routing paths in the WSN. If a huge proportion of sN are alive and accessible in the sensor network, a
large volume of data can be collected from the targeted region. The graph for such total number of alive nodes for
each round is shown in Figure 10. In LEACH and EHR-DC, the first node dies at 215 rounds, FLMFLA at 225 rounds,
and MSFLMFLA at 235 rounds. In comparison to FLMFLA, EHRDC, and LEACH,The sN in MSFLMFLA has
clearly drained away energy in a relatively equal way, and for even more rounds all the nodes have become alive, and
makes the MSFLMFLA more steady. In this scenario, MSFLMFLA outperformed its competitors because node
expiration occurs at future rounds in comparison with LEACH, EHRDC, and FLMFLA. As nodes expire prematurely
in every round, because of the reported inadequacy of field surveillance, the network might become inadequate. Nodes
might also die after just a prolonged period of time following their deployment to improve reliability and stability.

Fig 10:Number of Round vs alive node
Fig 11: Number of Round vs. dead nodes
8.3 Remaining Energy of the network
A sensor node wastes a massive quantity of energy in wireless communication. As that of the rounds grows, the
network's remaining network energy declines, resulting in death of the node. Remaining network energy is depicted
in Figure 12. In this scenario, we can see that the remaining energy of MSFLMFLA is always greater than that of
FLMFLA, EHRDC, and LEACH for round intervals of 10, 50, 100, 150, 200, and 250. It demonstrates that the
proposed method efficiently distributes the WSN's load, tends to result in a more stable region.

Fig 12: Remaining Energy of the network
8.4 Throughput of the Network
Throughput can be defined as over the course of a round sending more packets towards the BS, implying that as more
data is obtained out from target area.Figure 13 depicts the throughput of sensing data received by the base
station.When compared to FLMFLA, EHRDC and LEACH, MSFLMFLA delivered 5%, 10%, and 12% more packets
to BS.
Fig 13:Throughput of the Network
8.5 FND,LND and Throughput

When the network connection is established, the goal is to gather as much information as possible out from targeted
region, in which the node's life is critical. This is due to the fact that the dying of every node tends to make the network
vulnerable to non coverage of certain portions of the targeted region. As a result, the overall system design will perform
poorly. Figure 14, Figure 15 and Figure 16 depicts FND, LND and The simulation results provided throughput metrics.
In this scenario, MSFLMFLA's FND (Network stability) is protracted by 20%, 12%, and 4% when compared to the
LEACH, EHRDC, and FLMFLA protocols, respectively. Similarly, LND (Network Lifetime) is extended by 10%,
6.66%, and 3.333% when compared to LEACH, EHRDC and FLMFLA, respectively.Finally, in comparison to
LEACH, EHRDC, and FLMFLA.Throughput is increased by 12%, 10%, and 5%, respectively. The results obtained
from simulation show here that MSFLMFLA can meet the application's necessities while having a longer lifetime and
stability period.
Fig 14: Network Stability-First node dead
Fig 15: Network Lifetime-Last node dead

Fig 16:Throughput
9 Conclusion
Sensors have limited energy but must continue to report data via multiple relays, so efficient energy routing is critical
to extending WSN life. The suggested methodology enables an MDS to move even within the sensor deployed in the
network as well as tour sensors in most efficient manner, lowering path length. In network there by applications in
which its preasumed that information across all the sensors is important as equal, The MSFLMFLA decreases the
length of the tour for MDS. In terms of alive nodes, dead nodes, throughput, and residual energy, the results of
simulation demonstrate that the MSFLMFLA algorithm outperforms the LEACH, EHR-DC, and FLMFLA
algorithms.This demonstrates that the proposed algorithm is appropriate across all WSN situations and has a long
lifespan. As a result, MSFLMFLA not only increases network life, throughput, and reduces end-to-end delay, but it
also increases data received by the sink.
10 Declarations
10.1 Funding
we confirm that No funding was received for this work.
10.2 Conflicts of Interest/Competing interests
We wish to confirm that there are no known conflicts of interest associated with this publication and there has been
no significant financial support for this work that could have influenced its outcome.
10.3 Availability of Data and material
The data that support the findings of this study are openly available in “MSFLMFLA” and by all the authors name
and with the paper refference number and with its doi.
10.4 Code Availability
Not applicable.
10.5 Authors Contribution
All the work is done by Author 1 and Manuscipt is written by Author 1 under the guidence of Author 2 and Author 3.
All listed Authors meet the WPC criteria .We attest that all authors contributed significantly to the creation of this
manuscript each having fullfilled criateria as establisted by the WPC.

10.6 Ethics approval
As the journal is a member of the Committee on Publication Ethics (COPE) .We confirm that we have given due
consideration to the protection of intellectual property accociated with this work and that there are no impediments to
publication, including the timing of Publication,with respect to intellectual property .In so doing we confirm that we
have followed the regulations of our instituions concerning intellectual property.
10.7 Consent to participate
We confirm that the manuscript has been read and approved by all named authors and there are no other persons who
satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the
manuscript has been approved by all of us
10.8 Consent for Publication
Not applicable
References
1. H. Ennajari, Y. Ben Maissa, and S. Mouline, “Energy efficient in-network aggregation algorithms in wireless sensor
networks: A survey,” Lect. Notes Electr. Eng., vol. 397, pp. 135–148, 2017, doi: 10.1007/978-981-10-1627-1_11.
2. A. A. Baradaran and K. Navi, “CAST-WSN: The Presentation of New Clustering Algorithm Based on Steiner Tree
and C-Means Algorithm Improvement in Wireless Sensor Networks,” Wirel. Pers. Commun., vol. 97, no. 1, pp. 1323–
1344, 2017, doi: 10.1007/s11277-017-4572-x.
3. G. Han and L. Zhang, “WPO-EECRP: Energy-Efficient Clustering Routing Protocol Based on Weighting and
Parameter Optimization in WSN,” Wirel. Pers. Commun., vol. 98, no. 1, pp. 1171–1205, 2018, doi: 10.1007/s11277-
017-4914-8.
4. B. Baranidharan and B. Santhi, “DUCF: Distributed load balancing Unequal Clustering in wireless sensor networks
using Fuzzy approach,” Appl. Soft Comput. J., vol. 40, pp. 495–506, 2016, doi: 10.1016/j.asoc.2015.11.044.
5. R. Elhabyan, W. Shi, and M. St-Hilaire, “Coverage protocols for wireless sensor networks: Review and future
directions,” J. Commun. Networks, vol. 21, no. 1, pp. 45–60, 2019, doi: 10.1109/JCN.2019.000005.
6. Y. Zhang, M. Liu, and Q. Liu, “An energy-balanced clustering protocol based on an improved CFSFDP algorithm
for wireless sensor networks,” Sensors (Switzerland), vol. 18, no. 3, 2018, doi: 10.3390/s18030881.
7. Agrawal D., Pandey S. (2020) Load Balanced Fuzzy-Based Clustering for WSNs. In: Khanna A., Gupta D.,
Bhattacharyya S., Snasel V., Platos J., Hassanien A. (eds) International Conference on Innovative Computing and
Communications. Advances in Intelligent Systems and Computing, vol 1059. Springer, Singapore.
https://doi.org/10.1007/978-981-15-0324-5_49
8. V. Anand, D. Agrawal, P. Tirkey, and S. Pandey, “An Energy Efficient Approach to Extend Network Life Time of
Wireless Sensor Networks,” Procedia Comput. Sci., vol. 92, pp. 425–430, 2016, doi: 10.1016/j.procs.2016.07.332.
9. Y. Padmanaban and M. Muthukumarasamy, “Energy-efficient clustering algorithm for structured wireless sensor
networks,” IET Networks, vol. 7, no. 4, 2018, doi: 10.1049/iet-net.2017.0112.
10. S. Pandey, K. Dubey and S. sharma, "Performance improvement of LEACH protocol using Naïve Bayes Classifier
in Wireless Sensor Networks," 2021 International Conference on Computing, Communication, and Intelligent
Systems (ICCCIS), 2021, pp. 263-267, doi: 10.1109/ICCCIS51004.2021.9397209.
11. G. K. Nigam and C. Dabas, “ESO-LEACH: PSO based energy efficient clustering in LEACH,” J. King Saud Univ.
- Comput. Inf. Sci., 2018, doi: 10.1016/j.jksuci.2018.08.002.

12. V. Anand and S. Pandey, “Particle Swarm Optimization and harmony search based clustering and routing in
Wireless Sensor Networks,” Int. J. Comput. Intell. Syst., vol. 10, no. 1, p. 1252, 2017, doi: 10.2991/ijcis.10.1.84.
13. D. Agrawal and S. Pandey, “FUCA: Fuzzy-based unequal clustering algorithm to prolong the lifetime of wireless
sensor networks,” Int. J. Commun. Syst., vol. 31, no. 2, pp. 1–18, 2018, doi: 10.1002/dac.3448.
14. Y. Zhang, J. Wang, D. Han, H. Wu, and R. Zhou, “Fuzzy-logic based distributed energy-efficient clustering
algorithm for wireless sensor networks,” Sensors (Switzerland), vol. 17, no. 7, 2017, doi: 10.3390/s17071554.
15. C. Fu, Z. Jiang, W. Wei, and A. Wei, “An Energy Balanced Algorithm of LEACH Protocol in WSN,” Int. J.
Comput. Sci. Issues, vol. 10, no. 1, pp. 354–359, 2013.
16. S. Pandey, N. K. Nagwani, and C. Kumar, “Cluster based - SPIN routing protocol for wireless sensors networks,”
Indian J. Sci. Technol., vol. 8, no. 15, 2015, doi: 10.17485/ijst/2015/v8i15/49800.
17. L. Xiao, F. Wu, D. Yang, T. Zhang, and X. Zhu, “Energy Efficient Wireless Sensor Network Modelling Based on
Complex Networks,” J. Sensors, vol. 2016, 2016, doi: 10.1155/2016/3831810.
18. V. Anand and S. Pandey, “New approach of GA–PSO-based clustering and routing in wireless sensor networks,”
Int. J. Commun. Syst., vol. 33, no. 16, pp. 1–20, 2020, doi: 10.1002/dac.4571..
19. M. Baskaran and C. Sadagopan, “Synchronous Firefly Algorithm for Cluster Head Selection in WSN,” Sci. World
J., vol. 2015, 2015, doi: 10.1155/2015/780879.
20. X. S. Yang and X. He, “Firefly algorithm: recent advances and applications,” Int. J. Swarm Intell., vol. 1, no. 1,
p. 36, 2013, doi: 10.1504/ijsi.2013.055801.
21. V. Srivastava, S. Tripathi, K. Singh, and L. H. Son, “Energy efficient optimized rate based congestion control
routing in wireless sensor network,” J. Ambient Intell. Humaniz. Comput., vol. 11, no. 3, pp. 1325–1338, 2020, doi:
10.1007/s12652-019-01449-1.
22. M. Xu and G. Liu, “A multipopulation firefly algorithm for correlated data routing in underwater wireless sensor
networks,” Int. J. Distrib. Sens. Networks, vol. 2013, 2013, doi: 10.1155/2013/865154.
23. A. N. Shahbaz, H. Barati, and A. Barati, “Multipath routing through the firefly algorithm and fuzzy logic in
wireless sensor networks,” Peer-to-Peer Netw. Appl., vol. 14, no. 2, pp. 541–558, 2021, doi: 10.1007/s12083-020-
01004-2.
24. A. Hamzah, M. Shurman, O. Al-Jarrah, and E. Taqieddin, “Energy-efficient fuzzy-logic-based clustering
technique for hierarchical routing protocols in wireless sensor networks,” Sensors (Switzerland), vol. 19, no. 3, pp.
14–16, 2019, doi: 10.3390/s19030561.
25. R. M. Al-Kiyumi, C. H. Foh, S. Vural, P. Chatzimisios, and R. Tafazolli, “Fuzzy Logic-Based Routing Algorithm
for Lifetime Enhancement in Heterogeneous Wireless Sensor Networks,” IEEE Trans. Green Commun. Netw., vol. 2,
no. 2, pp. 517–532, 2018, doi: 10.1109/TGCN.2018.2799868.
26. H. Jiang, Y. Sun, R. Sun, and H. Xu, “Fuzzy-logic-based energy optimized routing for wireless sensor networks,”
Int. J. Distrib. Sens. Networks, vol. 2013, 2013, doi: 10.1155/2013/216561.
27. X. Zhao, S. Ren, H. Quan, and Q. Gao, “Routing protocol for heterogeneous wireless sensor networks based on a
modified grey wolf optimizer,” Sensors (Switzerland), vol. 20, no. 3, pp. 1–18, 2020, doi: 10.3390/s20030820.
28. B. Pitchaimanickam and G. Murugaboopathi, “A hybrid firefly algorithm with particle swarm optimization for
energy efficient optimal cluster head selection in wireless sensor networks,” Neural Comput. Appl., vol. 32, no. 12,
pp. 7709–7723, 2020, doi: 10.1007/s00521-019-04441-0.
29. A. Al-Baz and A. El-Sayed, “A new algorithm for cluster head selection in LEACH protocol for wireless sensor
networks,” Int. J. Commun. Syst., vol. 31, no. 1, pp. 1–13, 2018, doi: 10.1002/dac.3407.

30. I. Mosavvar and A. Ghaffari, “Data Aggregation in Wireless Sensor Networks Using Firefly Algorithm,” Wirel.
Pers. Commun., vol. 104, no. 1, pp. 307–324, 2019, doi: 10.1007/s11277-018-6021-x.
31. Z. Al Aghbari, A. M. Khedr, W. Osamy, I. Arif, and D. P. Agrawal, “Routing in Wireless Sensor Networks Using
Optimization Techniques: A Survey,” Wirel. Pers. Commun., vol. 111, no. 4, pp. 2407–2434, 2020, doi:
10.1007/s11277-019-06993-9.
32. P. S. Mehra, M. N. Doja, and B. Alam, “Fuzzy based enhanced cluster head selection (FBECS) for WSN,” J. King
Saud Univ. - Sci., vol. 32, no. 1, pp. 390–401, 2020, doi: 10.1016/j.jksus.2018.04.031.
33. J. Wang, Y. Gao, W. Liu, A. K. Sangaiah, and H. J. Kim, “An improved routing schema with special clustering
using PSO algorithm for heterogeneouswireless sensor network,” Sensors (Switzerland), vol. 19, no. 3, 2019, doi:
10.3390/s19030671.
34. S. Pandey and V. Anand, “Load-Balanced Clustering Scheme with Sink Mobility for Heterogeneous Wireless
Sensor Networks,” Natl. Acad. Sci. Lett., vol. 40, no. 5, pp. 335–341, 2017, doi: 10.1007/s40009-017-0590-1.
35. Y. C. Wang and B. C. Pan, “Event-aware hierarchical routing with differential compression to extend WSN
lifetime,” APNOMS 2020 - 2020 21st Asia-Pacific Netw. Oper. Manag. Symp. Towar. Serv. Netw. Intell.
Humanit.,pp.79–84,2020, doi:10.23919/APNOMS50412.2020.9237022.
36. G. Yogarajan and T. Revathi, “Nature inspired discrete firefly algorithm for optimal mobile data gathering in
wireless sensor networks,” Wirel. Networks, vol. 24, no. 8, pp. 2993–3007, 2018, doi: 10.1007/s11276-017-1517-y.
37. P. Yadav, V. Kishore Yadav, and S. Yadav, “Distributed Energy Efficient Clustering Algorithm to Optimal Cluster
Head by Using Biogeography Based Optimization,” Mater. Today Proc., vol. 5, no. 1, pp. 1545–1551, 2018, doi:
10.1016/j.matpr.2017.11.244.