Journal of the

Nigerian Society
of Physical
J. Nig. Soc. Phys. Sci. 4 (2022) 945
Sciences

Efficient and Intelligent Decision Support System for Smart

Irrigation
Monika Sainia , Ashish Kumara,∗, Vijay Singh Maana , Deepak Sinwarb
a Department of Mathematics & Statistics, Manipal University Jaipur, Jaipur-303007, India
b Department of computer and Communication, Manipal University Jaipur, Jaipur-303007, India

Abstract
The main aim of present analysis is to develop a novel efficient and intelligent irrigation system (EIIS). The proposed irrigation system configured
using five components arranged in a series configuration along with the internal cold standby redundancy on sensor unit. The failure and repair
rates are exponentially distributed. By using the Markovian birth-death process differential difference equations of the model are developed to
derive the availability expressions and estimation of parameters. The availability of the system is optimized by employing Grey-Wolf optimization
(GWO) and Dragon Fly algorithm (DA) for efficiency and performance evaluation. The derived results are helpful for the system designers.

DOI:10.46481/jnsps.2022.945

Keywords: Markov model, Availability, Cold standby redundancy, Intelligent irrigation system

Article History :
Received: 20 July 2022
Received in revised form: 28 September 2022
Accepted for publication: 14 October 2022
Published: 11 November 2022

c 2022 The Author(s). Published by the Nigerian Society of Physical Sciences under the terms of the Creative Commons Attribution 4.0 International license
(https://creativecommons.org/licenses/by/4.0). Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI.
Communicated by: T. Latunde

1. Introduction smart irrigation system based on soli and environmental mea-

surements irrigation decision can be made. Several researchers
Agriculture is one of the oldest occupations done by human like Morais et al. [1], Vellidis et al. [2], Kehui et al. [3], Giusti
being. It is one of the strong pillars that contribute to the de- and Marsili-Libelli [4], Navarro-hellı́n et al. [5], and Sinwar
velopment of the economic development of any country. A big et al. [6] proposed model for smart irrigation systems in var-
portion of the population of any country depends on the agricul- ious aspects. Nasikou et al. [7] proposed a model for smart
ture for survival. In most of the countries traditional methods energy utilization in smart irrigation systems. Gu et al. [8] de-
of irrigation are adopted for farming but these methods suffer veloped a software for irrigation scheduling that work on the
due to several drawbacks. The wasteful use of water and inap- crop water stress. Shekhar et al. [9] developed an automated
propriate irrigation are the major drawbacks while smart irriga- irrigation system using intelligent IoT. Goap et al. [10] used
tion technology ensures the efficient use of water in irrigation. machine learning and open source technologies in development
The traditional method of irrigation is very challenging due to of a internet based irrigation management systems. Nawandar
several factors like soil nature, crop requirement, etc. while in and Satpute [11] devloped a smart irrigation system based on
IoT . A lost cost intelligent module has beem utilized in the de-
velopment of the system. Wang et al. [12] proposed a decision
∗ Correspondingauthor tel. no: +917725922864
Email address: [email protected] (Ashish Kumar) support system to manage the canal irrigation. It is observed
1
 Saini et al. / J. Nig. Soc. Phys. Sci. 4 (2022) 945 2

that the performance aspects of these smart system has not been intelligent decisions for irrigation based on the data cap-
extensively explored so far. Reliability and availability are the tured by sensors in real time.
major concern with the perofrmance of these systems. Many
researchers worked in the direction of reliability evaluation and
2. Notations
optimization of performance of systems. Maihulla and Yusuf
[13] examined the reliability, availability, maintainability and The nomenclatures used in the development of the model are as
dependability to check the sensitivity effect in a grid-connected follows:
photovoltaic systems. This sensitivity analysis shows that close S i : ith state of smart irrigation system
attention and close monitoring are needed to ensure the dis- Pi (t): Probability that smart irrigation system is in state i at time
tribution board’s reliability of the system. Venkatakrishnan et t
0
al. [14] compare the effect of modified differential evolution Pi (t): Derivative of first order of Pi (t)
algorithm (MDE) on wind turbine system along with GA and A, B1 , B2 , C, D, E: Description of the operative states of smart
PSO techniques. It concludes that in addition to meet energy irrigation system
demand, the modified DE algorithm assists in finding the sys- a, b1 , b2 , c, d, e: Description of the failed states of smart irriga-
tem’s most cost-effective solution. Kumar et al. [15] proposed tion system
an efficient model for availability optimization of cooling tower αi /βi i = 1, 2, 3, 4, 5: Failure/repair rates of A, B1 , B2 , C, D, E
using metaheuristic algorithms. Saini et al. [16] optimized the respectively.
performance of a biological and chemical processing unit us-
ing genetic algorithm and particle swarm optimization. Saini et
3. System Description
al. [17] proposed a stochastic model for availability optimiza-
tion of condenser used in steam turbine power plants using GA In this section, the configuration of the efficient and intelli-
and PSO. Though area of reliability optimization of smart irri- gent irrigation system is presented. A smart irrigation system
gation systems is still untouched. But efficient and intelligent is presented in Figure 1 that depicts five main components viz.
irrigation-based systems are necessary nowadays for optimum power unit, active and cold standby sensor unit, Raspberry pi,
utilization of fresh water. So, the work is proposed develop such water pump, and irrigation unit. As the name suggest, power
an efficient and intelligent irrigation system which ensures the unit is needed to provide the power supply to the smart irriga-
availability of the system as and when required for irrigation. tion systems. The power may come directly or through solar
The rest of the work is arranged in six sections: Section power supply. On the other hand, sensor unit indicates a col-
2 devoted to the novelty claims by the investigators, section 3 lection of sensors (i.e., moisture, temperature, water, humidity,
explained the notations and section 4 briefed the system de- etc.) needed for decision making by the control unit. The heart
scription. Mathematical model proposed in section 5 and in of this smart irrigation system is Raspberry pi that is responsible
section 6 numerical discussion and graphical representation is for taking efficient irrigation decisions based on data captured
made followed by concluding section 7. by sensor unit. In addition, Raspberry pi is utilized to trigger
A very less efforts have been made so far for development the water pump whenever required. Once the water pump is
of efficient and intelligent irrigation systems. So, here an effort turned on, the irrigation unit (i.e., sprinklers) start working to-
is made to develop and efficient and intelligent irrigation sys- wards irrigation of the field. Upon reaching the threshold values
tem. The efficiency of the system is evaluated in terms of avail- by sensor units, Raspberry pi initiates actions to turn off the wa-
ability of the system for use. By probabilistic arguments and ter pump. It is evident that sensor unit is more prone to failures,
Markov methodology, a mathematical model is developed and that’s why the redundancy has been utilized in sensor unit of
optimized using Grey-Wolf optimization (GWO) and Dragon the system. If primary sensor failed in fetching information any
Fly algorithm (DA). The state transition diagram of proposed one or more of these than standby sensor starts immediately and
system is shown in Figure 2. The blow mentions points are failed sensor undergoes for repair. The concept of cold standby
claimed as the novelty by the inventors. redundancy and exponential distributed random variable have
been utilized in development of the stochastic model. The re-
1. Design: A novel design of the proposed system is de-
pair and switch devices are perfect and sufficient repair facility
veloped for efficiency improvement. The cold standby
is available with system. The architecture and state transition
redundancy is used at sensor level, and it is verified that
diagram of smart irrigation system is depicted in Figure 1 and
it performed better.
Figure 2 respectively.
2. Efficient: The availability of the proposed irrigation sys-
tem is optimized using various algorithms like Grey-Wolf 4. Mathematical Modelling and Analysis
optimization (GWO) and Dragon Fly algorithm (DA). In
literature so far these algorithms are not employed on Here, mathematical model for the smart irrigation system
availability optimization of irrigation systems. The sys- is developed using Markov birth-death process. The Chapman-
tem is proved efficient when it is operated with the pa- Kolmogorov differential difference equations derived based on
rameters estimated by GWO parameters. Figure 2.
3. Intelligent: The proposed system is capable of taking P0 (t+∆t) = (1−α1 ∆t−α2 ∆t−α3 ∆t−α4 ∆t−α5 ∆t) P0 (t)
2
 Saini et al. / J. Nig. Soc. Phys. Sci. 4 (2022) 945 3

Figure 1. Framework of intelligent decision support system for smart irrigation

Figure 2. State Transition Diagram

0
+β1 P1 (t) ∆t+β2 P2 (t) ∆t+β3 P3 (t) ∆t+β4 P4 (t) ∆t+β5 P5 (t)∆t P0 (t) = − (α1 +α2 +α3 +α4 +α5 ) P0 +β1 P1 (t) +β2 P2 (t)
P0 (t+∆t) −P0 (t)
lim = − (α1 +α2 +α3 +α4 +α5 ) P0 +β1 P1 (t) +β3 P3 (t) +β4 P4 (t) +β5 P5 (t)
∆t→0 ∆t
+β2 P2 (t) +β3 P3 (t) +β4 P4 (t) +β5 P5 (t)
3
 Saini et al. / J. Nig. Soc. Phys. Sci. 4 (2022) 945 4

we get
" ! !#−1
0
lim P0 (t) = − (α1 +α2 +α3 +α4 +α5 ) P0 +β1 P1 +β2 P2 α2 α1 α2 α3 α4 α5
t→∞
P0 = 1 + 1 + + + + + (15)
β2 β1 β2 β3 β4 β5
+β3 P3 +β4 P4 +β5 P5 − (α1 +α2 +α3 +α4 +α5 ) P0 +β1 P1 +β2 P2
The system availability function is defined as:
+β3 P3 +β4 P4 +β5 P5 = 0 (1)
A0 = P0 + P2
α1 P0
P1 = (2) !" ! !#−1
β1 α2 α2 α1 α2 α3 α4 α5
= 1+ 1+ 1+ + + + +
β2 β2 β1 β2 β3 β4 β5
β1 P6 +β2 P7 +β3 P8 +β4 P9 +β5 P10 + α2 P0
P2 = (3) (16)
α1 +α2 +α3 +α4 +α5 + β2

α3 P0 5. Numerical Results and Discussion

P3 = (4)
β3
In this section, parameter estimation of the failure and re-
α4 P0 pair rates of the smart irrigation system is done using swarm-
P4 = (5) intelligence based algorithms namely Gray Wolf Optimization
β4
(GWO) and Dragon Fly algorithm (DA). The possible search
α5 P0 space of the decision variables is appended in Table 1. The
P5 = (6)
β5 estimated values of the parameters with respect to 30 iteration
levels at several population sizes is derived and appended in
α1 P2 Table 2. The availability of smart irrigation system derived at
P6 = (7)
β1 various population sizes are appended in Figure 3-6. The exe-
cution time of the program taken by algorithms are appended in
α2 P2
P7 = (8) Table 6. The simulation study is performed using R software on
β2 Windows 10 64-bit operating system having 8 GB of RAM and
α3 P2 Intel Core i5 8th generation CPU. The range of these decision
P8 = (9) variables is provided in Table 1 as follows:
β3
Here parameters of failure and repair rates for GWO are
α4 P2 constant for different population sizes and system attains its
P9 = (10) maximum availability in early stage of simulation process. Pa-
β4
rameters values change rapidly when using DA algorithm. It is
α5 P2 seen that failure rate of Controller Arduino uno increase rapidly
P10 = (11)
β5 for population size 600 and water pump failure increase with
population size 400 and 1000.
The initial conditions are as follows: Table 3 appended the various parameters values after 50
iterations corresponding to various population sizes. It also
P0 (0) = 1
shows that by using GWO, parameters for failure rates and re-
Pi (0) = 0 where i = 1 to 10 (12) pair rates have its minimum and maximum value simultane-
ously and system attains its maximum availability. for DA,
The set of linear equations [1]-[11] along with initial con- range of parameters of failure rates increase for sub-system
ditions [12] constitute the mathematical model for smart irriga- sensor/standby sensor unit, Controller Arduino uno and water
tion system. After simplification the probabilities at respective pump for population size 800 and 1000.
states are derived as follows: Table 4 highlights the estimated values after 70 iterations on
different population sizes. For GWO, system attains its maxi-
α1 α2 α3 α4 mum availability. While using DA, range of parameters of fail-
P0 = P0 , P1 = P0 , P2 = P0 , P3 = P0 , P4 = P0 ,
β1 β2 β3 β4 ure rates increase rapidly for sub-system sensor/standby sensor
α5 α1 α2 α2 α2 α3 α2 unit for population size 800.
P5 = P0 , P6 = P0 , P7 = P0 , P8 = P0 ,
β5 β1 β2 β2 β2 β3 β2 Table 5 reported the estimated parametric values of failure
α4 α2 α5 α2 and repair rates after 90 iterations at various population sizes.
P9 = P0 , P10 = P0 (13)
β4 β2 β5 β2 For GWO, system attains its maximum availability. While us-
ing DA, range of parameters of failure rates increase rapidly
By normalizing criteria that sum of all are transition proba- for sub-system Controller Arduino uno for population size 600
bilities is equal to 1 and 1000. The best parameter values of failure and repair rates,
10
X derived by simulation process done in R studio for GWO and
Pi = 1 (14) DA optimizations. Parameters obtained for different popula-
i=0 tion sizes and various iterations. These are the best parameters
4
 Saini et al. / J. Nig. Soc. Phys. Sci. 4 (2022) 945 5

Table 1. Range of the decision variables

Sub-system Range of failure-rate (α) Range of repair-rate (β)
Power unit α1 = [0.000003, 1.99] β1 = [0.000009, 2.11]
Sensor unit &Stand by Sensor unit α2 = [0.000005, 1.89] β2 = [0.000006, 2.30]
Controller Arduino uno α3 = [0.000001, 1.53] β3 = [0.000008, 2.57]
Water pump α4 = [0.000004, 1.24] β4 = [0.000007, 2.08]
Smart valve α5 = [0.000002, 1.32] β5 = [0.000019, 2.34]

Table 2. Parameter estimation of various failure and repair rates after 30 iterations and different population sizes by using GWO and DA
Iter\NP 400 600 800 1000
GWO α1 0.0000030 0.0000030 0.0000030 0.0000030
α2 0.0000050 0.0000050 0.0000050 0.0000050
α3 0.0000010 0.0000010 0.0000010 0.0000010
α4 0.0000040 0.0000040 0.0000040 0.0000040
α5 0.0000020 0.0000020 0.0000020 0.0000020
β1 1.86464 2.11 2.11 2.11
β2 2.3 2.3 2.3 2.3
β3 2.57 2.57 2.57 2.57
β4 2.08 2.08 2.08 2.08
β5 2.34 2.34 2.34 2.34
DA α1 0.0000030 0.0000030 0.0000030 0.0000030
α2 0.0000050 0.0000050 0.243093 0.0000050
α3 0.0000010 1.53 0.0000010 0.0000010
α4 1.24 0.0000040 0.0000040 1.24
α5 0.0000020 0.0000020 0.0000020 0.0000020
β1 2.11 2.11 2.11 2.11
β2 0.6341965 2.118504 2.3 2.3
β3 2.57 2.57 1.130091 2.57
β4 2.08 2.08 2.08 2.08
β5 2.34 2.34 1.496042 2.34

Table 3. Parameter estimation of various failure and repair rates after 50 iterations and different population sizes by using GWO and DA
Iter\NP 400 600 800 1000
GWO α1 0.0000030 0.0000030 0.0000030 0.0000030
α2 0.0000050 0.0000050 0.0000050 0.0000050
α3 0.0000010 0.0000010 0.0000010 0.0000010
α4 0.0000040 0.0000040 0.0000040 0.0000040
α5 0.0000020 0.0000020 0.0000020 0.0000020
β1 2.11 2.11 2.11 2.11
β2 2.3 2.3 2.3 2.3
β3 2.57 2.57 2.57 2.57
β4 2.08 2.08 2.08 2.08
β5 2.34 2.34 2.34 2.34
DA α1 0.0000030 0.0000030 0.0000030 0.0000030
α2 0.0000050 0.0000050 1.047884 1.414718
α3 0.0000010 0.8860956 1.190195 0.02381295
α4 0.0000040 0.0000040 0.0000040 1.24
α5 0.0000020 0.0000020 0.2853557 0.0355639
β1 0.239885 2.11 0.7081718 2.11
β2 2.3 2.3 2.3 2.3
β3 0.6369218 1.704359 2.57 2.57
β4 2.08 0.1690187 2.08 2.08
β5 2.34 1.701183 2.34 2.34

5
 Saini et al. / J. Nig. Soc. Phys. Sci. 4 (2022) 945 6

Table 4. Parameter estimation of various failure and repair rates after 70 iterations and different population by using GWO, DA
Iter\NP 400 600 800 1000
GWO α1 0.0000030 0.0000030 0.0000030 0.0000030
α2 0.0000050 0.0000050 0.0000050 0.0000050
α3 0.0000010 0.0000010 0.0000010 0.0000010
α4 0.0000040 0.0000040 0.0000040 0.0000040
α5 0.0000020 0.0000020 0.0000020 0.0000020
β1 2.11 2.11 2.11 2.11
β2 2.3 2.3 2.3 2.3
β3 2.57 2.57 2.57 2.57
β4 2.08 2.08 2.08 2.08
β5 2.34 2.34 2.34 2.34
DA α1 0.0000030 0.0000030 0.0000030 0.0000030
α2 0.0000050 0.05620236 1.89 0.0000050
α3 0.0000010 0.0000010 0.0000010 0.0000010
α4 0.4775554 0.1183404 0.0000040 0.1965477
α5 0.0000020 0.8200995 0.0000020 0.0000020
β1 2.11 0.00640018 2.11 2.11
β2 0.1106202 2.3 2.3 2.3
β3 2.57 0.7383446 1.424582 2.267976
β4 1.718693 1.26267 2.08 0.8067474
β5 1.063321 2.34 2.34 2.34

Table 5. Parameter estimation of various failure and repair rates after 90 iterations and different population sizes by using GWO and DA
Iter\NP 400 600 800 1000
GWO α1 0.0000030 0.0000030 0.0000030 0.0000030
α2 0.0000050 0.0000050 0.0000050 0.0000050
α3 0.0000010 0.0000010 0.0000010 0.0000010
α4 0.0000040 0.0000040 0.0000040 0.0000040
α5 0.0000020 0.0000020 0.0000020 0.0000020
β1 2.11 2.11 2.11 2.11
β2 2.3 2.3 2.3 2.3
β3 2.57 2.57 2.57 2.57
β4 2.08 2.08 2.08 2.08
β5 2.34 2.34 2.34 2.34
DA α1 1.059087 0.0000030 0.5533655 0.0000030
α2 0.0000050 0.4501209 0.0000050 0.0000050
α3 0.6671488 1.53 0.0000010 1.53
α4 0.0000040 0.4808177 0.0000040 0.1965477
α5 0.0000020 0.1602266 0.0000020 0.0000020
β1 2.11 1.725929 2.11 2.11
β2 0.5978837 1.181086 2.3 2.3
β3 1.702016 2.57 2.57 2.57
β4 1.457232 1.438981 0.2638387 2.08
β5 2.34 2.34 2.34 2.34

Table 6. Elapsed time (in seconds) of the GWO and DA algorithms used in finding the optimum availability with respect to iterations at various population size
Iteration Population size
400 600 800 1000
GWO DA GWO DA GWO DA GWO DA
30 3.96 9.37 3.72 7.01 4.5 8.52 4.32 7.89
50 4.18 7.75 4.61 7.82 7.98 7.69 4.2 12.61
70 3.81 9.8 3.55 7.1 3.79 7.05 3.58 7.39
90 3.37 7.17 3.92 7.75 3.63 7.2 4.35 7.84

6
 Saini et al. / J. Nig. Soc. Phys. Sci. 4 (2022) 945 7

Figure 6. Availability of Smart Irrigation System at various iterations at Pop.

Figure 3. Availability of Smart Irrigation System at various iterations at Pop. Size =1000
Size =400

Availability for DA decreases rapidly from 0.9899925 to 0.5788153

in-between iteration 30 to 50. Then after it increases to 0.7922205
for iteration 90.
Figure 6 is the representation of the optimum availability
of system using GWO and DA with respect to various itera-
tions with population size 1000. System attains its maximum
availability 0.9999954 in early stage of population size and it-
eration by using GWO and then remain constant for higher pop-
ulation size and iterations. For DA availability varies between
Figure 4. Availability of Smart Irrigation System at various iterations at Pop.
Size =600 0.804096 at iteration 70 to 0.5391205 at iteration 50.
Table 6 show elapsed time taken by optimization techniques
which includes time taken by system and optimization tech-
nique and provide the total execution time. The variation in
time taken to complete the task by using GWO and DA can be
easily seen. Time taken by GWO is almost half of the time
taken by DA. It is also seen that GWO attains the maximum
availability at early stage with respect to DA.

Figure 5. Availability of Smart Irrigation System at various iterations at Pop. 6. Conclusion

Size =800
In present work, a novel efficient and intelligent irrigation
system is developed using the concept of cold standby redun-
for which system attains its maximum availability. For GWO,
dancy. The availability expression for the model is derived
parameters for failure rates and repair rates have its minimum
and optimized using the Grey-Wolf optimization (GWO) and
and maximum value simultaneously and system attains its max-
Dragon Fly (DA) algorithms. The parameters of all failure
imum availability at early stage of simulation. For DA the val-
and repair rates are estimated by both algorithms. The maxi-
ues of parameters are fluctuating rapidly, and sub-systems need
mum availability derived is 0.9999954 in the search space by
adequate maintenance at time.
both the algorithms. It is observed that Dragon fly algorithm is
In Figure 3, graphical representation of the optimum avail-
not given sustain results with the increase of number of itera-
ability of system using GWO and DA with respect to various
tions. Though the elapse time taken by Grey-Wolf optimization
iterations with population size 400 is shown. It is seen that at
(GWO) is sufficiently less in comparison to Dragon Fly (DA)
iteration 30 and 50 both techniques attain the same and max-
algorithm. So, it is recommended that the system when oper-
imum availability after that availability for DA is decreasing
ated according to the parameters estimated by grey wolf opti-
with increasing the iteration size and for GWO it remains same
mization perform more efficiently. The present work may be
at 0.9999954.
further extended to the component level investigation by using
In Figure 4, graphical representation of the optimum avail-
some other nature inspired algorithms. The concept of simulta-
ability of system using GWO and DA with respect to various it-
neous failure and redundancy can be involved in further study.
erations with population size 600 is shown. Here availability for
The proposed methodology may be opted in other process in-
GWO remains constant at 0.9999954 for each iteration. Avail-
dustries.
ability for DA slightly increase from 0.6268276 to 0.6919241
for iteration 30 and 70 respectively but iteration 90 availability
decreases rapidly to 0.4754869. Acknowledgement
Figure 5 shows the graphical representation of the optimum
availability of system using GWO and DA with respect to var- The authors are extremely grateful to the insightful com-
ious iterations with population size 800 is shown. Availabil- ments provided by the reviewers and section editor to improve
ity for GWO remains constant at 0.9999954 for each iteration. the quality of this study.
7
 Saini et al. / J. Nig. Soc. Phys. Sci. 4 (2022) 945 8

References [10] A. Goap, D. Sharma, A. K. Shukla & C. Rama Krishna, “An IoT
based smart irrigation management system using Machine learning and
[1] R. Morais, A. Valente & C. Serôdio, A wireless sensor network for smart open source technologies”, Computers and electronics in agriculture 155
irrigation and environmental monitoring: A position article. In 5th Euro- (2018) 41, https://doi.org/10.1016/j.compag.2018.09.040
pean federation for information technology in agriculture, food and envi- [11] N. K. Nawandar & V. R. Satpute, “IoT based low cost and intelligent
ronment and 3rd world congress on computers in agriculture and natural module for smart irrigation system”, Computers and electronics in agri-
resources (EFITA/WCCA) (July, 2005) 845. culture 162 (2019) 979, https://doi.org/10.1016/j.compag.2019.05.027
[2] G. Vellidis, M. Tucker, C. Perry, C. Kvien & C. Bednarz, “A real-time [12] W. Wang, Y. Cui, Y. Luo, Z. Li & J. Tan, “Web-based decision support
wireless smart sensor array for scheduling irrigation”, Computers and system for canal irrigation management”, Computers and Electronics in
electronics in agriculture 61 (2008) 44. Agriculture 161 (2019) 312.
[3] X. Kehui, X. Deqin & L. Xiwen, “Smart water-saving irrigation system [13] A. S. Maihulla, & I. Yusuf, “Performance analysis of photovoltaic sys-
in precision agriculture based on wireless sensor network”, Transactions tems using (RAMD) analysis”, Journal of the Nigerian Society of Physi-
of the Chinese society of Agricultural Engineering 26 (2010) 170. cal Sciences 3 (2021) 172.
[4] E. Giusti & S. Marsili-Libelli, “A Fuzzy Decision Support System for ir- [14] G. R. Venkatakrishnan, R. Rengaraj, K. K. Sathish, R. K. Dinesh & T.
rigation and water conservation in agriculture”, Environmental Modelling Nishanth, “Implementation of modified differential evolution algorithm
& Software 63 (2015) 73. for hybrid renewable energy system”, Journal of the Nigerian Society of
[5] H. Navarro-hellı́n, J. Martı́nez-del-rincon, R. Domingo-miguel, F. Soto- Physical Sciences 3 (2021) 209.
valles & R. Torres-sánchez, “A decision support system for managing [15] A. Kumar, M. Saini, N. Gupta, D. Sinwar, D. Singh, M. Kaur, & H. N.
irrigation in agriculture”, Computers and Electronics in Agriculture 124 Lee, “Efficient stochastic model for operational availability optimization
(2016) 121. of cooling tower using metaheuristic algorithms”, IEEE Access 10 (2022)
[6] D. Sinwar, V. S. Dhaka, M. K. Sharma & G. Rani, “AI-based yield pre- 24659.
diction and smart irrigation”, Internet of Things and Analytics for Agri- [16] M. Saini, D. Goyal, A. Kumar, & R. B. Patil, “Availability optimization of
culture 2 (2020) 155. biological and chemical processing unit using genetic algorithm and par-
[7] A. Nasiakou, M. Vavalis & D. Zimeris, “Smart energy for smart irriga- ticle swarm optimization”, International Journal of Quality & Reliability
tion”, Computers and Electronics in Agriculture 129 (2016) 74. Management 39 (2022) 1704.
[8] Z. Gu, Q. Zhiming, M. Liwang, G. Dongwei, X. Junzeng, F. Quanxiao, [17] M. Saini, N. Gupta, V. G. Shankar & A. Kumar, “Stochastic modeling and
Y. Shouqi & F. Gary, “Development of an irrigation scheduling software availability optimization of condenser used in steam turbine power plants
based on model predicted crop water stress”, Computers and Electronics using GA and PSO”, Quality and Reliability Engineering International 38
in Agriculture 143 (2017) 208. (2022) 2670.
[9] Y. Shekhar, E. Dagur, S. Mishra, R. J. Tom, M. Veeramanikandan & S.
Sankaranarayanan, “Intelligent IoT based automated irrigation system,”
International Journal of Applied Engineering Research 12 (2017) 7306.