IAES International Journal of Artificial Intelligence (IJ-AI)

Vol. 13, No. 2, June 2024, pp. 2265~2274

ISSN: 2252-8938, DOI: 10.11591/ijai.v13.i2.pp2265-2274  2265

A new optimal strategy for energy minimization in wireless

sensor networks

Hicham Ouchitachen1, Anouar Darif1, Mohamed Er-Rouidi2, Mustapha Johri3

1
Laboratory of Innovation in Mathematics, Applications and Information Technologies (LIMATI), Faculty of Polydisciplinary,
Sultan Moulay Slimane University, Beni Mellal, Morocco
2
Department of Mathematics and Computer Science, Modeling and Combinatorial Laboratory, Faculty of Polydisciplinary,
Cadi Ayyad University, Safi, Morocco
3
ENSAM, Mohammed V University, Rabat, Morocco

Article Info ABSTRACT

Article history: In recent years, evolutionary and metaheuristic algorithms have emerged as
crucial tools for optimization in the field of artificial intelligence. These
Received Jul 24, 2023 algorithms have the potential to revolutionize various aspects of our lives by
Revised Dec 1, 2023 leveraging the multidisciplinary nature of wireless sensor networks (WSNs).
Accepted Dec 29, 2023 This study aims to introduce genetic and simulated annealing algorithms as
effective solutions for enhancing WSN performance. Our contribution
entails two main phases. Firstly, we establish mathematical models and
Keywords: formulate objectives as a nonlinear constrained optimization problem.
Secondly, we develop two algorithmic solutions to address the formulated
Energy consumption optimization problem. The obtained results from multiple simulations
Metaheuristic algorithms demonstrate the positive impact of the proposed strategies on improving
Multi-objective genetic network performance in terms of energy consumption.
algorithm
Optimization problem This is an open access article under the CC BY-SA license.
Wireless sensor networks

Corresponding Author:
Hicham Ouchitachen
Laboratory of Innovation in Mathematics, Applications and Information Technologies (LIMATI)
Faculty of Polydisciplinary, Sultan Moulay Slimane University
Beni Mellal, Morocco
Email: [email protected]

1. INTRODUCTION
Recently, optimal resource management in wireless sensor networks (WSNs) has gained significant
importance and is considered a fertile domain for numerous research studies [1], [2]. To support multiple
applications simultaneously over a single physical WSN architecture, it is crucial to design and utilize WSNs
in an optimal manner, considering their limited computing, memory, and battery power [3], [4]. In this
context, ubiquitous sensor networks, which incorporate WSN technologies and internet of things (IoT)
paradigms, are rapidly advancing to sense unpredictable environments and provide support to remote clients
[5]–[7]. The collected data are processed and trans’
mitted to a cloud system, leveraging the processing and storage capabilities of the resource-
constrained devices [8], [9]. Given that most real-time applications in WSNs require substantial resource
improvements [10], [11], it is worth mentioning that WSNs consist of numerous sensors deployed in diverse
environments to gather data on various environmental phenomena. These networks find applications in
earthquake monitoring, marine activity detection, fire detection, smart grids, and intelligent transportation
systems [12]. Furthermore, WSNs are extensively used in industrial, medical, environmental, agricultural,
and home automation applications [13]–[16].
In WSN, sensor nodes can be deployed in the target area either randomly or deterministically,
depending on their applications [17]. These different applications are based on optimizing multiple

Journal homepage: http://ijai.iaescore.com

2266  ISSN: 2252-8938

objectives, which can include bandwidth utilization, storage capacity, and computational capabilities of the
sensors. This optimization aims to enhance overall network efficiency, extend sensor lifetime, ensure reliable
data transfer, and reduce network maintenance costs. Additionally, optimal resource management facilitates
accurate data collection by minimizing packet losses and reducing interference. It also contributes to
optimizing the quality of services provided by WSNs, guaranteeing continuous, and reliable monitoring of
environmental phenomena. Overall, these resources depend significantly on the energy of sensors. In other
words, one of the primary challenges of WSNs is managing energy constraints. Since sensors are typically
powered by batteries, efficient management of this limited resource is essential. Excessive energy
consumption can significantly reduce sensor lifetime and necessitate frequent battery replacements, which
can be costly and impractical. Optimal deployment of nodes, especially the base station (BS) relocation, is
considered one of the research issues strongly related to energy constraints in WSNs. Indeed, depending on
the chosen communication technology, data transmission may not be guaranteed if the distance between the
transmitter and receiver is too large. This challenge is particularly prominent in applications involving
surveillance of large areas and deployments with very low density.
The study conducted in this paper aims to minimize energy consumption in WSNs using BS
relocation to optimize network performance. This approach has two key advantages: i) it reduces sensor
energy consumption and ii) as a direct result of this reduction, it significantly extends the network's lifetime.
Our contribution focuses on the development of two optimal strategies: the optimal strategy based on
simulated annealing (OSSA) strategy and the optimal strategy based on multi-objective genetic algorithms
(OSGA) strategy. This study is connected to various prior works, offering a contextual foundation that
emphasizes the significance of energy constraints in WSNs. It reexamines certain mathematical models
employed for energy optimization in these networks and references recent research endeavors presenting
algorithmic solutions aimed at tackling this issue. In the literature, several research works have focused on
the study of optimal resource management in WSNs [18]. Given that the wireless communication module
consumes the most energy, the primary objective of these optimization techniques is to minimize the radio
module's activity time for different nodes. These techniques can be grouped into two classes: those based on
data reduction and those based on sleep/active mode switching.
Yin et al. [19] proposed an optimization method based on the Yin-Yang pigeon-inspired
optimization algorithm aimed at guaranteeing optimal management of a WSN. To optimize the coverage
requirements in WSNs, Elhoseny et al. [20] introduced a model that utilized the genetic algorithm for
continuous monitoring of specific targets with limited energy resources. Chowdhury et al. [21] presented an
energy-efficient optimization technique using the Voronoi-glowworm swarm optimization-K-means
algorithm. Jebi and Baulkani [22] proposed a multi-objective randomized grasshopper optimization
algorithm-based selective activation method for optimal WSN management. Ali and Özdağ [23] elaborated
on a new metaheuristic approach using the elfes probabilistic detection model to optimize WSN performance.
Zulfiqar et al. [24] introduced an optimization algorithm to effectively manage a WSN in terms of energy
consumption. Abdulzahra et al. [25] presented with the aim of minimizing operational costs and emissions
while considering the variability of energy sources.
According to the literature review presented in this section, it is clear that the reduction of energy
consumption is a significant research issue in WSNs. Specifically, energy constraints are a critical concern
for WSNs deployed in hostile or inaccessible areas and networks with high density. In these scenarios,
energy resources are severely limited, making it challenging or even impossible to recharge or replace sensor
batteries after deployment. The objective of this paper is to address the energy limitations encountered by
WSNs. Specifically, this study aims to enhance sensor performance in terms of efficient communication with
the BS, thereby extending the network's lifetime. Our contribution consists of two main phases: phase 1 is
mathematical modeling, focusing on formulating the objectives to be achieved as a nonlinear constrained
optimization problem. Phase 2 is algorithmic resolution, concentrating on the development of two optimal
multi-objective strategies to solve the problem formulated in phase 1. These two phases will be presented in
section 2.
Our research methodology unfolds as follows: in the first step, we identify limitations in the existing
literature regarding energy constraints in WSNs. In the second step, we address the aforementioned issues
using the proposed approach. This approach involves the development of two optimal strategies for BS
relocation to improve network performance in terms of energy consumption. These strategies are based on
constrained nonlinear optimization algorithms, namely genetic algorithms and simulated annealing. In the
third step, we demonstrate and validate the positive impact of the proposed algorithms through simulation
results. The rest of this paper is organized as follows: section 2 elaborate on the various phases of the
proposed approach; section 3 present the obtained results through multiple simulations, these results are
described, analyzed, and interpreted; and section 4 provide concluding remarks and discuss future works.

Int J Artif Intell, Vol. 13, No. 2, June 2024: 2265-2274

Int J Artif Intell ISSN: 2252-8938  2267

2. PROPOSED APPROACH
2.1. Optimization problem formulation
In this study, we consider a WSN which is deployed in a defined area. The goal is to contribute to
reducing energy consumption in the network, thereby enhancing its performance and extending its
operational lifetime. Specifically, our approach aims to optimize sensor performance in terms of energy while
using two new optimal relocation strategies to determine the best placement for the BS. To achieve this, we
assign a function, denoted as 𝑒𝑖 , to each sensor 𝑠𝑖 to represent its energy consumption. Based on the
Heinzelman model, the energy consumed by a sensor 𝑠𝑖 when transmitting a message of 𝑙 bits over a distance
𝑑𝑖 in free space can be expressed as in (1),

𝑒𝑖 = 𝑙 × 𝐸𝑒𝑙𝑒𝑐 + 𝑙 × 𝑑𝑖2 × 𝐸𝑓𝑠 (1)

Where 𝐸𝑒𝑙𝑒𝑐 is the electric energy in 𝑝𝐽/𝑏𝑖𝑡, 𝐸𝑓𝑠 is the free space energy in 𝑝𝐽/𝑏𝑖𝑡/𝑚2 , 𝑙 is the message size
in 𝑏𝑖𝑡, and 𝑑𝑖 is the transmission distance between sensor 𝑠𝑖 and the BS in 𝑚.
The distance 𝑑𝑖 between sensor 𝑠𝑖 and the BS can be calculated using their coordinates (𝑥𝑖 − 𝑦𝑖 ) and
(𝑥𝑏 − 𝑦𝑏 ) respectively, based on the Euclidean distance as in (3),

𝑑𝑖 = √(𝑥𝑖 − 𝑥𝑏 )2 + (𝑦𝑖 − 𝑦𝑏 )2 (2)

Therefore,

𝑒𝑖 (𝑥𝑏 − 𝑦𝑏 ) = 𝑘 × 𝐸𝑒𝑙𝑒𝑐 + 𝑘 × ((𝑥𝑖 − 𝑥𝑏 )2 + (𝑦𝑖 − 𝑦𝑏 )2 ) × 𝐸𝑓𝑠 (3)

Minimizing the energy 𝑒𝑖 ensures an improvement in the performance of sensor 𝑠𝑖 . As shown in (3), the
value of this energy largely depends on the position of sensor 𝑠𝑖 relative to the BS. Therefore, a better
placement of the BS will positively influence the energy of the sensors and subsequently enhance the
network performance. In other words, the objective of this study can be formulated as a key question as
follows: 'how to determine an optimal position of the BS that minimizes the energy consumed by the various
sensors to improve network performance? Using the previously mentioned notations, we can formulate this
objective mathematically as a nonlinear constrained multi-objective optimization problem, as in (4),

𝑚𝑖𝑛𝐹(𝑥𝑏 − 𝑦𝑏 ) = (𝑒1 (𝑥𝑏 − 𝑦𝑏 ), 𝑒2 (𝑥𝑏 − 𝑦𝑏 ), 𝑒3 (𝑥𝑏 − 𝑦𝑏 ),..., 𝑒𝑛 (𝑥𝑏 − 𝑦𝑏 )) (4)

𝑠𝑡 𝐴 ≤ 𝑥𝑏 ≤ 𝐵
𝐶 ≤ 𝑦𝑏 ≤ 𝐷

Where 𝑛 is a number of sensors. [𝐴, 𝐵] × [𝐶, 𝐷] represents the area where the sensors are deployed. The
formulated problem (4) is a constrained nonlinear optimization problem. To solve this problem by
determining the optimal position of the BS, we propose two optimal strategies: OSSA strategy and OSGA
strategy, which will be elaborated in the next subsection.

2.2. Developed optimal strategies

2.2.1. Optimal strategy based on simulated annealing strategy
a. OSSA principle
The OSSA strategy utilizes simulated annealing, a meta-heuristic optimization method. It is
employed to solve combinatorial optimization problems, particularly when dealing with a large number of
sensors. The objective is to identify the best solution among numerous possibilities. Simulated annealing
operates by probabilistically exploring the search space, accepting suboptimal solutions in order to escape
local minima and discover the global optimum. The technique involves constructing a representation of the
problem's solution space, where each point is referred to as a state. Transitions between states occur by
making small adjustments to the current solution. To control the convergence of the OSSA strategy towards
𝑜𝑝 𝑜𝑝
the optimal solution (𝑥𝑏 , 𝑦𝑏 ), which represents the best position for the BS, a control parameter denoted as
𝑃𝑘 is utilized. This parameter determines the probability of accepting a less favorable solution compared to
the current one. In this study, we consider the following three schemes for updating the value of 𝑃𝑘 at
iteration k: i) exponential update: 𝑃𝑘 = 0,95𝑘 , ii) logarithmic update: 𝑃𝑘 = 1/log(𝑘), and iii) linear update:
𝑃𝑘 = 1/𝑘
b. OSSA objective function formulation
The OSSA strategy introduces multi-objective optimization to jointly minimize several objective
functions. For this, OSSA uses the weighted summation technique to form the objective function to be
minimized. Indeed, according to the modeling phase detailed in the previous section, each sensor 𝑠𝑖 is
characterized by a function 𝑒𝑖 which measures its energy. In the context of OSSA's multi-objective
A new optimal strategies for energy minimization in wireless sensor networks (Hicham Ouchitachen)
2268  ISSN: 2252-8938

optimization technique, the function 𝑒𝑖 represents the ith objective function that OSSA aims to minimize. In a
network with 𝑛 sensors, there are 𝑛 objective functions to optimize. OSSA employs the weighted summation
technique to combine these functions and formulate a single objective function F, as indicated in (5),

𝐻(𝑥𝑏 − 𝑦𝑏 ) = 𝛼1 𝑒1 (𝑥𝑏 − 𝑦𝑏 ) + 𝛼2 𝑒2 (𝑥𝑏 − 𝑦𝑏 ) + ... + α𝑛 𝑒𝑛 (𝑥𝑏 − 𝑦𝑏 ) (5)

where 𝛼𝑖 is the weight of the objective function 𝑒𝑖 . These weights satisfy the condition at (6):

𝛼1 + 𝛼2 + ... + α𝑛 = 1 (6)

c. OSSA flowchart
To achieve the optimal solution, the OSSA algorithm follows a sequence of procedural steps. The
initial step entails the initialization of input parameters (𝑆0 : initial solution and 𝑃0 : control parameter), where
the initial solution is regarded as the current solution, denoted as 𝑆𝑐𝑢𝑟 . Subsequently, the second step
generates a new solution 𝑆𝑛𝑒𝑤 , based on the current solution. In the third step, we evaluate the objective
function 𝐹 for both 𝑆𝑛𝑒𝑤 and 𝑆𝑐𝑢𝑟 , then we compute the difference 𝛥𝐹 = 𝐹(𝑆𝑛𝑒𝑤 ) − 𝐹(𝑆𝑐𝑢𝑟 ). If 𝛥𝐹 ≤ 0, the
subsequent step involves the generation of a random number 𝑟 between 0 and 1, then the computation of
exp(𝛥𝐹/𝑃𝑘 ). If 𝑟 < exp(𝛥𝐹/𝑃𝑘 ), 𝑆𝑐𝑢𝑟 value is preserved, and 𝑃𝑘 value is updated. If 𝛥𝐹 > 0, we accept
𝑆𝑛𝑒𝑤 as the current solution and verify if the stopping criterion is reached in order to determine the optimal
solution. The different steps of the OSSA strategy are summarized in the flowchart presented in Figure 1.

Figure 1. OSSA flowchart

2.2.2. Optimal strategy based on multi-objective genetic algorithms strategy

a. OSGA principle
The OSGA strategy aims to optimize multiple objectives simultaneously by incorporating the
principles of multi-objective optimization. This strategy is based on the concepts of Pareto dominance and
the search for non-dominated solutions. Pareto dominance is a fundamental concept in the design of OSGA.
It enables the comparison and ranking of solutions based on their performance across multiple objectives. In

Int J Artif Intell, Vol. 13, No. 2, June 2024: 2265-2274

Int J Artif Intell ISSN: 2252-8938  2269

the context of multi-objective optimization, a solution A is considered dominant over a solution B if it is at

least as good as B in all objectives and strictly better in at least one objective. In other words, A dominates B
if it outperforms B in at least one objective without being worse in any other objective. The basic idea behind
Pareto dominance is to search for solutions that cannot be improved simultaneously in all objectives. Instead
of seeking a single optimal solution, the focus is on finding a set of non-dominated solutions that represent
optimal trade-offs between the objectives. Following this principle, OSSA aims to identify a set of
non-dominated solutions known as the Pareto front. In this set, no solution can be improved in one objective
without compromising performance in another objective. This approach enables the OSGA strategy to rank
solutions based on their performance across multiple objectives, promoting the emergence of a set of
non-dominated solutions that represent the optimal compromises between the objectives.
b. OSGA objective function formulation
In order to optimize multiple objectives simultaneously, the OSGA strategy utilizes a fundamental
technique that treats these objectives as components of a single objective function represented as a vector.
Based on the mathematical modeling presented in section 2.1, each sensor 𝑠𝑖 is defined by a function 𝑒𝑖 that
represents its consumed energy. In the case of a network comprising 𝑛 sensors, employing the technique
introduced by OSGA, the objective function to be minimized can be expressed as (7),

𝐺(𝑥𝑏 − 𝑦𝑏 ) = [𝑒1 (𝑥𝑏 − 𝑦𝑏 ), 𝑒1 (𝑥𝑏 − 𝑦𝑏 ),..., 𝑒𝑛 (𝑥𝑏 − 𝑦𝑏 )] (7)

c. OSGA flowchart
In order to compute the optimal solution, the OSGA algorithm employs several steps. The first step
is to generate a list of initial solutions. To select the best solution, the second step evaluates the objective
function for the previously generated initial solutions. In the third step, we check if the chosen solution is
optimal according to the Pareto dominance concept. For each computed solution, there are two cases. If the
Pareto dominance condition is not met, another list of solutions is generated, and the search space is explored
by combining some features of these solutions. To prevent convergence to suboptimal solutions, some
potential solutions are randomly modified. If the Pareto dominance condition is met, then the last computed
solution is considered as the optimal solution. The various steps of OSGA are presented in Figure 2.

Figure 2. OSGA flowchart

A new optimal strategies for energy minimization in wireless sensor networks (Hicham Ouchitachen)
2270  ISSN: 2252-8938

3. SIMULATIONS AND RESULTS

To validate the proposed approach, numerous simulations were carried out using the following
parameters: Eelec = 50 pJ/bit, Efs = 10 pJ/bit/m2, the network area = 300×200 m2, and l = 4,000 bit. As an
initial configuration, we considered 10 sensors deployed within a surface area of 140×140. These sensors are
connected to a BS initially positioned at the center of the designated area. Although this conventional BS
placement strategy ensures complete coverage of the area, it incurs high energy consumption for the sensors.
This energy consumption primarily depends on the distance between the sensors and the BS. To address this
issue, we conducted a BS relocation using the OSSA optimal strategy. Figure 3 illustrates the results obtained
after implementing this optimal relocation. In these results, the value of Pk is updated according to the
exponential model.

6000
5000
E n e rg y ( m J)

4000
3000
2000
1000
0
S 1 S 2 S 3 S 4 S 5 S 6 S 7 S 8 S 9 S 10
S en sors

Figure 3. Consumption energy after relocatin BS using OSSA strategy in the initial configuration

3.1. Performances evaluation

3.1.1. Optimal strategy based on simulated annealing strategy
As an initial configuration, we consider 10 sensors deployed in a surface area of 140×140. These
sensors are connected to BS which is initially positioned at the center of the designated area. Although this
conventional BS placement strategy ensures complete coverage of the area, it incurs high energy
consumption for the sensors. This energy primarily depends on the distance between the sensors and the BS.
To address this, we conducted a BS relocation using the OSSA optimal strategy. Figure 3 shows the obtained
results after performing this optimal relocation. In these results, the value of 𝑃𝑘 is updated according to the
exponential model.
To evaluate the OSSA performance, we examine multiple configurations by varying the number of
utilized sensors. For each configuration, we compute the total energy consumption in the network after
relocating the BS using the OSSA strategy. Then compare this energy consumption to that incurred when the
BS is positioned at the center of the designated area using the conventional strategy. The obtained results are
shown in Figure 4. This figure demonstrates that the OSSA strategy has a positive influence on energy
consumption, thanks to the relocation of BS to its optimal position. To control the convergence of the OSSA
strategy to the optimal solution (best position of the BS), we conducted several simulations by updating the
values of the parameter 𝑃𝑘 using the following models: exponential (exp), logarithmic (log), and linear (lin).
The results provided by these simulations are presented in Figure 5. According to these results, it can be seen
that selecting the exponential scheme enables the OSSA strategy to converge rapidly to the optimal solution,
compared to the linear and logarithmic schems. This significant finding favors the exp scheme as it facilitates
the control parameter to converge quickly to 0. Consequently, the OSSA strategy can promptly reach the
stopping criterion.

3.1.2. Optimal strategy based on multi-objective genetic algorithms strategy

Considering the same initial configuration presented in the previous subsection, we relocated the BS
using the OSGA strategy, with the aim of minimizing energy consumption. Figure 6 illustrates the energy
consumption levels of the various sensors (S1, S2, ..., S10) following this optimal relocation. Based on the
conducted simulations, it has been observed that the OSGA convergence is influenced by the selection of an
appropriate value for a critical parameter known as the Pareto fraction (PF). Indeed, to choose PF optimal
value in terms of rapid convergence, we performed several simulations by varying the value of PF from 0.1
to 0.9 using the previous configurations. The obtained results are presented in Figure 7. This figure

Int J Artif Intell, Vol. 13, No. 2, June 2024: 2265-2274

Int J Artif Intell ISSN: 2252-8938  2271

demonstrates that the optimal value of PF, which ensures a rapid convergence of the OSGA strategy, is 0.5.
Due to its importance in terms of convergence, in the rest of this paper, all remaining simulations will be
performed using this optimal value.
In order to evaluate the efficiency of the optimal OSGA strategy compared to the classical strategy,
we investigate several configurations with varying numbers of sensors. For each configuration, we quantify
the total energy consumption by the sensors relative to the two aforementioned strategies. The obtained
results are presented in Figure 8. This figure clearly shows the significant reduction in total energy
consumption achieved by the second optimal strategy proposed in this paper, namely the OSGA strategy.
This reduction is attributed to the optimal positioning of the BS, which leverages the advantageous properties
of multi-objective optimization.

Log Lin Exp

600000

500000 OSSA C la s s ic s tr a te g y 7000

C o n s u m e d e n e rg y ( m J)

6000

Iteration convergence
400000
5000
300000 4000
3000
200000
2000
100000 1000
0
0
10 20 30 40 10 20 30 40
Nu m b er of sen sors Nu m b er of sen sors

Figure 4. Consumption energy: OSSA compared to Figure 5. OSSA convergence relatively to 𝑃𝑘

classic strategy update schems

8000 200
7000 180
160
C o n v e r g e n c e ite r a tio n

6000
E nerg y (m J)

140
5000
120
4000 100
3000 80
2000 60
40
1000
20
0 0
S 1 S 2 S 3 S 4 S 5 S 6 S 7 S 8 S 9 S 10 0 0 ,1 0 ,2 0 ,3 0 ,4 0 ,5 0 ,6 0 ,7 0 ,8 0 ,9 1

S e ns ors PF v alu e

Figure 6. Consumption energy after relocatin BS Figure 7. OSGA convergence according to the PF
using OSGA strategy in the initial configuration value

600000
C o n s u m e d e n e r g y (m J)

500000 OS G A c la s s ic s tra te g y

400000

300000

200000

100000

0
10 20 30 40

Num be r of s e ns ors

Figure 8. Consumption energy: OSGA compared to classic strategy

A new optimal strategies for energy minimization in wireless sensor networks (Hicham Ouchitachen)
2272  ISSN: 2252-8938

3.2. Comparison between OSSA and OSGA

3.2.1. Convergence to the optimal solution
Using various configurations, we conducted numerous simulations to examine the convergence of
OSSA and OSGA. the obtained results are illustrated in Figure 9. This figure shows that the OSGA strategy
exhibits rapid convergence to the optimal solution compared to OSSA. This advantage can be attributed to
the nature of the problem addressed in this paper, which involves the simultaneous optimization of multiple
objectives. We particularly emphasize the benefits of the OSGA strategy in formulating the objective
function, which facilitates parallelization of the optimization process. This parallelization significantly
reduces execution time and positively influences the convergence of the OSGA strategy.

3.2.2. Energy consumption

Using multiple configurations, we calculate the total energy consumed in the network. The obtained
results are presented in Figure 10. It shows that relocating the BS using the OSSA strategy results in
relatively lower energy consumption compared to the OSGA strategy. This is because the OSSA strategy
prioritizes the overall network by minimizing total energy consumption, rather than focusing on individual
sensor energy. Consequently, while using this strategy, some sensors may have high energy consumption and
others low consumption, but the total energy consumption remains minimal. Conversely, the OSGA strategy
primarily focuses on optimizing the energy consumed by each sensor individually. In other words, it achieves
a balanced total energy consumption among the different sensors by simultaneously optimizing the objective
functions. It is important to emphasize that in these critical scenarios, where certain sensors exhibit high
energy consumption leading to battery depletion and a decrease in the network's lifetime, achieving energy
balance in the network using the OSGA strategy is considered an effective solution to address such issues.

OSGA OSSA

2000
1800
C o n v e r g e n c e ite r a tio n

1600
1400
1200
1000
800
600
400
200
0
0 10 20 30 40

Num be r of s e ns ors

Figure 9. Comparison between OSSA and OSGA in terms of convergence

450000

400000
Energy consumption (mJ)

OSGA OSSA
350000

300000

250000

200000

150000

100000

50000

0
10 20 30 40
Num be r of s e ns ors

Figure 10. Comparison between OSSA and OSGA in terms of energy consumption

4. CONCLUSION
In this paper, we have presented two new optimal strategies for enhancing the performance of WSN
with a focus on energy consumption minimization. These strategies are based on the multi-objective genetic
algorithm and simulated annealing algorithm. To elaborate the proposed approach, the objectives were

Int J Artif Intell, Vol. 13, No. 2, June 2024: 2265-2274

Int J Artif Intell ISSN: 2252-8938  2273

initially formulated as a nonlinear constrained optimization problem. Following that, two algorithmic
solutions were developed to address the formulated problem. The obtained results have demonstrated the
significance and effectiveness of the proposed strategies in reducing energy consumption within the network.
This energy constraint is considered as one of the major research challenges in WSN. The future work will
focus on the integration of machine learning techniques and data analytics to enhance the accuracy of
predicting network behavior and optimize energy consumption in a more intelligent and proactive manner.

REFERENCES
[1] H. Ouchitachen, A. Hair, and N. Idrissi, “Improved multi-objective weighted clustering algorithm in Wireless Sensor Network,”
Egyptian Informatics Journal, vol. 18, no. 1, pp. 45–54, 2017, doi: 10.1016/j.eij.2016.06.001.
[2] A. Darif and H. Ouchitachen, “Performance Improvement of a New Mac Protocol for Ultra Wide Band Wireless Sensor
Networks,” Journal of Theoretical and Applied Information Technology, vol. 100, no. 4, pp. 1015–1026, 2022.
[3] R. Sridhar and N. Guruprasad, “Energy efficient chaotic whale optimization technique for data gathering in wireless sensor
network,” International Journal of Electrical and Computer Engineering, vol. 10, no. 4, pp. 4176–4188, 2020, doi:
10.11591/ijece.v10i4.pp4176-4188.
[4] S. Savitha, S. C. Lingareddy, and S. Chitnis, “Energy efficient clustering and routing optimization model for maximizing lifetime
of wireless sensor network,” International Journal of Electrical and Computer Engineering, vol. 10, no. 5, pp. 4798–4808, 2020,
doi: 10.11591/ijece.v10i5.pp4798-4808.
[5] K. Lakshmanna, N. Subramani, Y. Alotaibi, S. Alghamdi, O. I. Khalafand, and A. K. Nanda, “Improved Metaheuristic-Driven
Energy-Aware Cluster-Based Routing Scheme for IoT-Assisted Wireless Sensor Networks,” Sustainability, vol. 14, no. 13,
pp. 1–19, 2022, doi: 10.3390/su14137712.
[6] A. H. Wheeb and M. T. Naser, “Simulation based comparison of routing protocols in wireless multihop ad hoc networks,”
International Journal of Electrical and Computer Engineering, vol. 11, no. 4, pp. 3186–3192, 2021, doi:
10.11591/ijece.v11i4.pp3186-3192.
[7] M. A. A. da Cruz, J. J. P. C. Rodrigues, P. Lorenz, V. V. Korotaev, and V. H. C. de Albuquerque, “In.IoT—A New Middleware
for Internet of Things,” IEEE Internet of Things Journal, vol. 8, no. 10, pp. 7902–7911, 2021, doi: 10.1109/JIOT.2020.3041699.
[8] K. Haseeb, A. Almogren, I. U. Din, N. Islam, and A. Altameem, “SASC: Secure and authentication-based sensor cloud
architecture for intelligent internet of things,” Sensors, vol. 20, no. 9, pp. 1–18, 2020, doi: 10.3390/s20092468.
[9] A. Caggiano, “Cloud-based manufacturing process monitoring for smart diagnosis services,” International Journal of Computer
Integrated Manufacturing, vol. 31, no. 7, pp. 612–623, 2018, doi: 10.1080/0951192X.2018.1425552.
[10] F. Al-Turjman and H. Zahmatkesh, “A Comprehensive Review on the Use of AI in UAV Communications: Enabling
Technologies, Applications, and Challenges,” in Unmanned System Technologies, Cham: Springer, 2020, pp. 1–26, doi:
10.1007/978-3-030-38712-9_1.
[11] V. K. Quy, V. H. Nam, D. M. Linh, N. T. Ban, and N. D. Han, “Communication Solutions for Vehicle Ad-hoc Network in Smart
Cities Environment: A Comprehensive Survey,” Wireless Personal Communications, vol. 122, no. 3, pp. 2791–2815, 2022, doi:
10.1007/s11277-021-09030-w.
[12] S. Bhattacharjee, P. Roy, S. Ghosh, S. Misra, and M. S. Obaidat, “Wireless sensor network-based fire detection, alarming,
monitoring and prevention system for Bord-and-Pillar coal mines,” Journal of Systems and Software, vol. 85, no. 3, pp. 571–581,
2012, doi: 10.1016/j.jss.2011.09.015.
[13] A. Salim, W. Osamy, A. Aziz, and A. M. Khedr, “SEEDGT: Secure and energy efficient data gathering technique for IoT
applications based WSNs,” Journal of Network and Computer Applications, vol. 202, 2022, doi: 10.1016/j.jnca.2022.103353.
[14] M. Suguna and S. Sathiyabama, “Shift invariant deep convolution neural learning for resource efficient healthcare data
transmission in WSN,” Measurement: Sensors, vol. 25, pp. 1–7, 2023, doi: 10.1016/j.measen.2022.100627.
[15] G. S. Uthayakumar, B. Jackson, C. Ramesh Babu Durai, A. Kalaimani, S. Sargunavathi, and S. Kamatchi, “Systematically
efficiency enabled energy usage method for an IOT based WSN environment,” Measurement: Sensors, vol. 25, pp. 1–6, 2023,
doi: 10.1016/j.measen.2022.100615.
[16] L. Kaur and R. Kaur, “A survey on energy efficient routing techniques in WSNs focusing IoT applications and enhancing fog
computing paradigm,” Global Transitions Proceedings, vol. 2, no. 2, pp. 520–529, 2021, doi: 10.1016/j.gltp.2021.08.001.
[17] S. M. Mohamed, H. S. Hamza, and I. A. Saroit, “Coverage in mobile wireless sensor networks (M-WSN): A survey,” Computer
Communications, vol. 110, pp. 133–150, 2017, doi: 10.1016/j.comcom.2017.06.010.
[18] G. Anastasi, M. Conti, M. Di Francesco, and A. Passarella, “Energy conservation in wireless sensor networks: A survey,” Ad Hoc
Networks, vol. 7, no. 3, pp. 537–568, 2009, doi: 10.1016/j.adhoc.2008.06.003.
[19] J. Yin, N. Deng, and J. Zhang, “Wireless Sensor Network coverage optimization based on Yin–Yang pigeon-inspired
optimization algorithm for Internet of Things,” Internet of Things, vol. 19, pp. 1–13, 2022, doi: 10.1016/j.iot.2022.100546.
[20] M. Elhoseny, A. Tharwat, X. Yuan, and A. E. Hassanien, “Optimizing K-coverage of mobile WSNs,” Expert Systems with
Applications, vol. 92, pp. 142–153, 2018, doi: 10.1016/j.eswa.2017.09.008.
[21] A. Chowdhury, D. De, A. Raychaudhuri, and M. Bakshi, “Corrigendum to Energy-Efficient Coverage Optimization in Wireless
Sensor Networks based on Voronoi-Glowworm Swarm Optimization-K-means algorithm [AdHoc Networks Volume 122
(November 2021) 102660],” Ad Hoc Networks, vol. 139, 2023, doi: 10.1016/j.adhoc.2022.102907.
[22] R. C. Jebi and S. Baulkani, “Mitigation of coverage and connectivity issues in wireless sensor network by multi-objective
randomized grasshopper optimization based selective activation scheme,” Sustainable Computing: Informatics and Systems,
vol. 35, pp. 1–13, 2022, doi: 10.1016/j.suscom.2022.100728.
[23] N. H. Ali and R. Özdağ, “Coverage analysis and a new metaheuristic approach using the Elfes Probabilistic detection model in
Wireless sensor networks,” Measurement, vol. 200, 2022, doi: 10.1016/j.measurement.2022.111627.
[24] R. Zulfiqar, T. Javid, Z. A. Ali, and V. Uddin, “Novel metaheuristic routing algorithm with optimized energy and enhanced
coverage for WSNs,” Ad Hoc Networks, vol. 144, 2023, doi: 10.1016/j.adhoc.2023.103133.
[25] A. M. K. Abdulzahra, A. K. M. Al-Qurabat, and S. A. Abdulzahra, “Optimizing energy consumption in WSN-based IoT using
unequal clustering and sleep scheduling methods,” Internet of Things, vol. 22, 2023, doi: 10.1016/j.iot.2023.100765.

A new optimal strategies for energy minimization in wireless sensor networks (Hicham Ouchitachen)
2274  ISSN: 2252-8938

BIOGRAPHIES OF AUTHORS

Hicham Ouchitachen received his Bachelor of Science (B.S.) in Ingénierie

Mathématique and Informatique (IMI) from Faculty of Sciences and Technology (FST) at
Sultan Moulay Slimane University, Beni Mellal, Morocco in 2010. He obtained his Master of
Science (M.S.) in Informatique, Signaux and Télécommunications from Faculty of Sciences at
Mohammed V-Agdal University, Rabat, Morocco in 2012. He received the Ph.D. degree in
Computer Sciences from Faculty of Sciences and Technology of Beni Mellal in 2017. In 2019,
he become an assistant professor in Faculty of Multidisciplinary at Sultan Moulay Slimane
University Beni Mellal, Morocco. His research interests include wireless sensor networks,
mobile edge computing (MEC), internet of things, performance evaluation using advanced
techniques in game theory, genetic algorithms, MDP, and resource management in wireless
mobile networks. He is an active reviewer of various international conferences and journals.
He can be contacted at email: [email protected].

Anouar Darif received the bachelor in informatique electrothecnique,

electronique and automatique (IEEA) from Dhar El Mahraz Faculty of Sciences at Mohamed
Ben Abdellah University Fez, Morocco in 2005. He received the Diplôme d’Etudes Supérieurs
Approfondies in Computer Sciences and Telecommunications from Faculty of Sciences Rabat
in 2007. He received the Ph.D. degree in Computer Sciences and Telecommunications from
Faculty of Sciences of Rabat in 2015. He is currently an research and a teaching associate in
Faculty of Multidisciplinary at University of Sultan Moulay Slimane Beni Mellal, Morocco.
His research interests include wireless sensor network, mobile edge computing (MEC),
internet of things, cloud computing, and neural networks. He is an active reviewer of various
international conferences and journals. He can be contacted at email:
[email protected].

Mohamed Er-Rouidi is a dedicated researcher and a teacher-researcher at the

Multidisciplinary Faculty of Cady Aayad University in Safi, Morocco. Holding a Ph.D. in
Computer Science from the Faculty of Sciences and Techniques of Beni Mellal in 2019, he
developed a strong interest in ad hoc networks. His research work primarily focuses on
enhancing routing protocols and minimizing energy consumption in mobile ad hoc networks.
His areas of expertise include mobile ad hoc networks as well as the internet of things (IoT),
ever-evolving fields that require a deep understanding to address contemporary challenges. In
addition to his active research pursuits, he is actively involved as an evaluator for various
international conferences and journals, thus contributing to the ongoing advancement of these
fields. He can be contacted at email: [email protected].

Mustapha Johri is an associate professor in computer science and a doctor in

Applied Mathematics, he is a lecturer at ENSAM in Rabat, a research member of the M2CS
(Applied Mathematics and Informatics) Laboratory, Mohammed V University Rabat. In 2019,
he completed a doctoral thesis on the theoretical and numerical study of an inverse flow
problem in a porous medium. His research interests include inverse problems, wireless sensor
networks, machine learning, and numerical simulation of complex systems. He is an active
reviewer of various international conferences and journals. He can be contacted at email:
[email protected].

Int J Artif Intell, Vol. 13, No. 2, June 2024: 2265-2274