A comprehensive review on various optimization routing algorithms in VANET

Full text

A comprehensive review on various
optimization routing algorithms in VANET
J. Sumadeep
SRM Institute of Science
and Technology
sj4440@srmist.edu.in
K. Vadivukkarasi
SRM Institute of Science
and Technology
vadivukk@srmist.edu.in
R. Dayana
SRM Institute of Science
and Technology
dayanar@srmist.edu.in
P. Malarvezhi
SRM Institute of Science
and Technology
malarvip@srmist.edu.in
Abstract— These days safe and collision-free traveling
is possible due to the evolution of self-driving technology.
Self-driving or autonomous vehicles (AVs) can replace
human-operated cars. It is indeed maintaining
communication between vehicles, infrastructure, and
pedestrians. The dynamic nature of nodes in vehicular ad-
hoc networks (VANET) creates a major challenge in
disseminating the data to a destination node by various
routing algorithms. This paper focused mainly on the
survey of the Meta-heuristic algorithm for routing
decisions.
Keywords- Meta-heuristics, Routing protocol,
Routing algorithm, VANET, Autonomous vehicles,
Optimization.
I. INTRODUCTION
Every year because of road accidents nearly 1.3
million people are dying which grabs many researchers'
interests in ensuring safe and collision-free
traveling. Vehicular ad hoc networks are an emerging
field in the automobile industry that leads to the
evolution of autonomous vehicles. VANET follows the
principles of mobile ad hoc networks (MANET and
provides a distinguished approach for Intelligent
Transport System (ITS) that enable different types of
communication services i.e., between Vehicles (V2V),
between Vehicle and Infrastructure (V2I), between
Infrastructure and Vehicle (I2V), and between Vehicle
and Everything (V2X) to provide road safety apart from
that, it also provides many value-added services like
video/audio sharing, broadcasting, and other
multimedia applications.
In recent years, various companies in the
automobile industry like Tesla, Waymo, BMW, Ford,
Audi, Toyota, etc., have been competing with each
other to release a full self-driving (FSD). Already
people from various countries are using semi-
autonomous cars and companies are trying very hard to
get the first fully self-driving car as early as possible.
The first self-sufficient and truly autonomous cars
appeared in the 1980s [1].
In 2021 Ford invests $7 billion in self-driving
cars [2]. Google sibling company Waymo which is
a leading manufacture company of autonomous
vehicles announced 2.5 billion dollars in March
2020 [3]. In 2016, to obtain a cruise automation GM
paid 581 million dollars [4]. In [5] in the year 2019
Uber raised $1billion on driverless cars from their
investors. Volvo in a joint venture with Uber also
invests $300 million in 2016. There are some non-
autonomous companies like Apple, Microsoft,
Cisco, Baidu, and Amazon which are also working
on developing fully autonomous vehicles. In [6]
according to McKinsey’s report during the period
2017-19, the car companies invested $120 billion.
Even though the technology in this sector is
booming, still there are few crashes of autonomous
cars were registering somewhere [7]. Carmakers
should provide safe and better driving assistance.
Figure 1 represents the estimated percentage of
automation vehicles.
Autonomous vehicles require both the
technological framework and the right legal to come
on roads. AVs will become a key feature for
transportation networks in the future and they are
categorized into 6 levels. They are (i) Level 0
vehicles, the driver must do everything (ii) Level 1
vehicles are driver assistance cars that control
steering. (iii) Level 2 vehicles are partial driving
automation vehicles that can control both
accelerating/deaccelerating and steering. (iv) Level
3 vehicles are conditional driving automation that
can operate two or more functions. (v) Level 4
vehicles are high driving automation which does not
require driver’s attention. (vi) Level 5 vehicles are
2022 First International Conference on Artificial Intelligence Trends and Pattern Recognition (ICAITPR) | 978-1-6654-2521-6/22/$31.00 ©2022 IEEE | DOI: 10.1109/ICAITPR51569.2022.9844205
Authorized licensed use limited to: SRM University. Downloaded on August 11,2022 at 10:11:18 UTC from IEEE Xplore. Restrictions apply.

100%
80%
2010- Improves
data collection and
performance
requirements for
autonomous
vehicles operating
on public roadways.
2019- Fully
autonomous vehicles
can drive from one
place to another
without any
interaction from the
driver.
2030- Self-driving
vehicles will be used
for various services.
2035- Globally, in a
year 12 million fully
autonomous
vehicles could be
sold.
2040- 75% of
vehicles will be
autonomous.
2050- supposed
reduce 90% of
accidents/crashes.
60%
2015- A short-range
fully driverless
vehicle service in
California was
launched by
Google.
2020- 10 million
self-driving vehicles
will be on the
highways.
2060-
Human driving may
be restricted.
40%
20%
2017- Autonomous
long-haul highway
trucks started
testing in Japan,
U.S, and Europe.
2025- Self-driving
features could
represent a
$42billion market.
0%
2010 2020 2030 2040 2050 2060
Figure 1: Estimated percentage of AVs
TABLE I. AUTOMATION VEHICLE LEVELS
Levels
L0
L1
L2
L3
L4
L5
Driver
Driver will
be the in-
charge
Driving must
do with basic
help
Must stay alert
Must be always ready to
take over within a
specified time or period
Can be a passenger,
but need to take
over driving when
the self-driving
system is unable to
continue
Driver not
required
Vehicle
Manual
control
Provide basic
help i.e.,
Lane-keeping,
automatic
emergency
braking
The vehicle can
perform steering,
braking, and
acceleration
Human override is still
required
Without any human
interaction, the
vehicle can perform
all the tasks under
special
circumstances
Fully automatic
human
interaction no
need
Driverless and can control vehicles completely
automatically. A lot of research is going on still in Level
5. Table 1 represents the different automation vehicle
levels. However, AVs are equipped with high-resolution
digital cameras and sensors to perceive their
surroundings and make sure, one should feel free to
travel. The remaining portion of the paper is
systematized as follows: Section 2 explains VANET
routing protocols. Section 3 describes various routing
algorithms based on metaheuristics in VANETs. Section
4 describes the effects of performance metrics in
VANETs and Section 5 concludes the paper.
II. ROUTING PROTOCOLS
In the early days, the drivers communicated frequently
using their voice, horns, hand gestures, and observation.
Later, in the 19th century, the usage of vehicles increases
drastically which led traffic police took a charge to control
the traffic by using gestures, colored lights, and semaphores.
In the year 1868 in London, the first-ever traffic lights were
installed which was operated manually, and then during the
Authorized licensed use limited to: SRM University. Downloaded on August 11,2022 at 10:11:18 UTC from IEEE Xplore. Restrictions apply.

1930s, the automation of traffic lights came into use. In the
year 1940s car makers deploy light indicators to the vehicles
to communicate. To adapt to the current circumstance for
driver’s various signs were introduced during the 1960s [8].
VANETs have been used in different scenarios
including traffic management, vehicle safety, driver
assistance, mobile entertainment, and emergency
broadcasting services. Communication between vehicles
and infrastructure is emerged rapidly by exchanging
information. The dissemination of information between
nodes plays a crucial role in the VANETs’ routing.
Routing in VANET is categorized into various types
based on topology, position, clustering, geo-cast, and
broadcasting [9] [10]. In VANET designing a dynamic
routing protocol is a principal challenge. While designing
a routing protocol one should consider the features like the
density of vehicles, change in topology, and varying
speed.
The various characteristics of VANETs are dynamic
topology, high mobility, dynamic network density, and
frequent exchange of information. From the last few
decades, certain routing algorithms are getting popular in
solving problems in the shortest period. Various
optimization issues are difficult to solve within a
reasonable time [11]. In those cases, the approximate
algorithm is the best choice to solve those problems. They
are broadly categorized into two: heuristic and
metaheuristic. Heuristics means to discover or enable
something and are specially designed to solve a specific
problem.
The main purpose of using a Metaheuristic is to find
all the possible solutions in a short period. Metaheuristics
are problem independent. They will be used to handle
bigger algorithms and solve problems faster in various
fields which leads them to use mostly for various issues.
This paper concentrates on various routing algorithms like
Metaheuristic.
III. METAHEURISTIC ROUTING ALGORITHMS
FOR VANET
Metaheuristics are broad categories into four. These
are based on Evolutionary, Nature, Trajectory, and
Ancient [12]. Figure 2 represents the categorization of
various metaheuristic algorithms.
A. Evolutionary based Metaheuristic
These models are inspired based on biological
evolution that includes mutation, reproduction, selection,
and recombination. To optimize the problem of the
population, considered candidates as a solution in
Evolutionary algorithms (EA) [13].
Figure 2. Metaheuristic Algorithms Categorization
The usage of EA is high because of their
performance for various types of problems and, they do
not make any assumptions ideally. In EA the
populations are generated randomly, and every
individual behaves like a solution. The appropriate
solution will be determined by an objective function
according to the pattern of choice at each stage, the
selection will be done based on better suitability and
high probability and then they will be regenerated by
various operators using mutation and crossover.
Some of the best techniques based on evolutionary
are Memetic Algorithm (MA) [14], Genetic Algorithm
(GA) [15], Harmony Search (HS) [16], Differential
Evolution (DE) [17], and Clonal Selection Algorithm
(CSA) [18] so on.
1. Genetic Algorithm (GA)
In GA, chromosomes were generated randomly
and assumed as population. Among them, the fittest
will survive and will be used for reproduction using
mutation and crossover. The chromosomes are called
genes. The genetic algorithm technique is used to solve
both constrained and unconstrained optimization
problems [22]. The steps involved in GA were listed
below.
Step 1: Chromosomes are generated randomly and
assumed as population.
Step 2: Among the population, the fittest one will
survive and will be used for reproduction.
Step 3: Generation of new chromosomes occurred
because of mating between parental chromosomes.
Metaheuristic
Evolutionary
Genetic algorithm
Evolutionary Strategies
Harmony Search..,
Trajectory
Guided Local Search
Variable
Neighbourhood Search
Tabu Search..,
Nature
Particle Swarm
Optimization
Plant based
Human based..,
Ancient Giza Pyramid
Construction
Authorized licensed use limited to: SRM University. Downloaded on August 11,2022 at 10:11:18 UTC from IEEE Xplore. Restrictions apply.

Step 4: Mutation maintains a genetic diversity between
old and new populations.
Step 5: Fitness will be calculated if it is ok the process
will end here else the process will repeat from step 2 to
step 4.
Joao F.M. Sarubbi et al. in [47] proposed a Delta-
GA to reduce the deployment of Road Side Units
(RSUs). Guoan Zhang et al. proposed a QoS perception
protocol based on a genetic algorithm and succeeded to
improve the packet’s transmission delay and packet
loss rate [48].
2. Memetic Algorithm (MA)
In [26], Dawkins developed a memetic algorithm,
which is like GA, here rather than genes the
chromosomes are called memes. This technique is
inspired based on local search behavior. The steps
involved in MA were listed below.
Step 1: Memes are generated randomly and assumed
as population.
Step 2: Among the population, the fittest one will
survive and will be used for reproduction using
equation 1.
Minimize F (T (s, D)) = w1Cost (T (s, D)) + w2 Delay
(R (s, D)) + w3 Jitter (R (s, D)) + w4 bandwidth (R (s,
D)) (1)
Step 3: New generation will be generated by
performing crossover and mutation like GA.
Step 4: perform the local search technique at last.
Step 5: Fitness will be calculated if it is ok the process
will end here else the process will repeat from step 2.
Marcelo F. Faraj et al. proposed a Memetic
algorithm based on gamma deployment metric to
deploy the RSUs and achieves almost 32.7% less
deployment compared to various algorithms [49].
3. Harmony Search (HS)
Harmony Search is a type of meta-heuristic
optimization technique, introduced by Zoong woo
Geem [27] inspired by the musician behavior. It can
balance exploitation and exploration. The steps
involved in Harmonic Search were listed below.
Step 1: Prepare a Harmony memory.
Step 2: Generate a new harmony randomly.
Step 3: If new > old, includes it in the New Harmony
memory.
Step 4: Else repeat Step 2 and Step 3.
Ravie Chandren Muniyandi et al. proposed an
improved harmony search algorithm by considering
OLSR parameters in highway scenarios and achieved
to improve the packet delivery ratio and less overhead
compared to existing algorithms [50].
4. Differential Evolution (DE)
It optimizes a problem based on an evolutionary
process by making repetition to improve candidate
solutions [17]. The steps involved in DE were listed
below.
Step 1: Generate the population weight vectors.
Step 2: Calculate the cost of each vector.
Step 3: Run from X=1 to NP
- Define the Xth vector in the
population.
- Create a new vector.
- Calculate the cost of the new vector.
- Compare new and old vectors and
choose the best.
Step 4: Check the criteria are satisfied or not. If not,
repeat from Step 2.
Step 5: If the criteria are satisfied then show the best
solution.
Er. Jayant Vasu et al. proposed a secure routing
protocol based on evolution to secure end-to-end (E2E)
communication between VANETs and RSUs [51].
5. Clonal Selection Algorithm (CSA)
This method is inspired by Darwin’s theory. It helps
to solve the various optimization problems by
mimicking the principles of clonal selection. The steps
involved in CSA were listed below.
Step 1: Generate the population
Step 2: Select some best individuals based on an
affinity measure among the population.
Step 3: The individuals are then cloned and become a
temporary population.
Step 4: The clones’ population further submitted to
hypermutation. Where it is proportional to their
affinity. Which helps to generate a matured antibody
population.
Step 5: The individuals from the mature antibody
population were selected to create a memory set. A set
of candidate solutions can be replaced by matured
antibody population.
Step 6: Antibodies are replaced by original ones.
Authorized licensed use limited to: SRM University. Downloaded on August 11,2022 at 10:11:18 UTC from IEEE Xplore. Restrictions apply.

In [52] Jinnah Hu et al. proposed a distribution
scheme in VANETs by considering both elastic and
inelastic types of contents based on a CSA which
reduces the delay time required and percentage of
failure for elastic and inelastic requests significantly.
B. Trajectory based Metaheuristic
This model is iteratively designed to work on a
single solution through paths. The trajectory-based
model starts with a random solution and looks to
improve the solution through searching the current
neighbourhood [19].
Some of the best techniques based on Trajectory are
Tabu Search (TS) [20] and Simulated Annealing (SA)
[21], so on.
1. Tabu Search (TS)
Tabu Search algorithm depends on the local search
procedure to solve the optimization problems
mathematically. TS further improves the quality of
local search by following its basic rules [28]. The steps
involved in Tabu Search is as follows.
Step 1: Select an initial solution x in T. Set x*=x and
z=0.
Step 2: Set z=z+1 and initialize a subset S* of solution
in N(x, z).
Step 3: Select the best y in S* and set X=y.
Step 4: If f'(x) < f(x*) then set x=x.
Step 5: Update condition.
Step 6: If the termination condition is satisfied then
stop. Else repeat from step 2.
Elham M et al, proposed a routing protocol that is
multi-level, based on Tabu search algorithm to improve
the QoS parameters such as link failure reduction,
packet loss reduction between vehicles, increase in
PDR while decreasing E2E delay and number of packet
losses [53].
2. Simulated Annealing (SA)
SA algorithm mimics the Physical Annealing
process. SA compares the outputs of the objective
functions repeatedly running with current and
neighboring nodes in the domain. If the neighboring
nodes generate a better result, then the next iteration is
saved as the base solution. Otherwise, the algorithm
terminates the procedure without searching. The steps
involved in SA were listed below.
Step 1: Set the system at an initial state Sk and compute
the function E(Sk).
Step 2: Choose another state Sk+1 and calculate the
function E(Sk+1).
Step 3: If E(Sk+1) ≤ E(Sk) then accept the transition to
the new state Sk = Sk+1 else with exp(−Δ𝐸
T).
Step 4: reduce the temperature by a cooling schedule.
Step 5: Repeat from Step 2 until global equilibrium.
Hosein Bagherlou et al. proposed a cluster-based
routing algorithm by considering simulated annealing
and achieving the best results in terms of packet
delivery ratio and route discovery rate [54].
C. Nature-based Metaheuristic
This method follows the rules and laws of nature to
find the optimum solution for a problem. Some of the
best techniques which are inspired by nature are Bio-
inspired, Swarm inspire, Plant-based,
Physics/Chemistry based Human-based methods.
Some of these algorithms were used by researchers in
VANET applications and were listed in Table 2 and
Table 3.
1. Swarm inspired Metaheuristic
A group of agents that work together can be defined
as a swarm. Swarm intelligence (SI) has been used in
various fields. SI is used to solve complex
mathematical problems. Some of the best examples of
the swarm are bees, fish, ants, birds, and termites.
There are many optimization algorithms in this
category such as Particle Swarm Optimization (PSO)
[29] [35], Ant Colony (ACO) [30] [43], Firefly (FA)
[31] [46], Artificial Bee Colony (ABC) [32] [33], so
on. Swarm intelligence is a subset of bio-inspired
algorithms. In Table 2, we regrouped the various
VANET routing algorithms with their merits and
demerits.
2. Bio-inspired Metaheuristic
Bio-inspired algorithms are used to solve complex
problems. These models follow the principles of the
biological behaviors of birds and animals to solve
problems. Some of the examples for bio-inspired
metaheuristic are Whale Optimization Algorithm
(WOA) [41], Fish Swarm Algorithm (FSA) [34], so on.
D. Ancient based Metaheuristic
Giza Pyramid Construction (GPC) is the first
algorithm based on ancient construction. In the future
based on historical monument constructions, many
algorithms may come but still, now no researcher used
this algorithm in any of the VANET applications. In
this method, the ancient past is the new origin for
inspiration. In the ancient past, there are various man-
made constructions which have numerous limitations.
Authorized licensed use limited to: SRM University. Downloaded on August 11,2022 at 10:11:18 UTC from IEEE Xplore. Restrictions apply.

The GPC is inspired by the ancient past [13]. There are
three large pyramids in the Giza pyramid complex site
also known as Giza Necropolis [23]. Many believed the
blocks used in the construction of pyramids were
brought by slaves from nearby mines, transported, and
then positioned in a place. It is difficult to place such
huge stone blocks, so they used ramps to place the
blocks at top levels [24] [25].
This algorithm depends upon the motion of workers
and stone blocks on the ramp. This technique has the
capability of handling high-dimensional problems. By
studying various constructional designs from past and
present civilizations one can design ancient, based
algorithms for routing decisions. The displacement of
stone blocks can be obtained using the following
equation. The steps that are taken while designing the
GPC were listed below.
𝑑 =
2
V
0
2𝑔(sin θ+ μk 𝑐𝑜𝑠θ) (2)
Step 1: Initialize the population size of workers and
blocks.
Step 2: Initialize position and cost of workers and
blocks.
Step 3: Determine best among workers as Pharaoh’s
agent.
Step 4: for i = 1...n do maximum iterations.
Step 5: Calculate the displacement of stone blocks.
Step 6: Calculate the movement of workers.
Step 7: Estimate the position of workers and blocks.
Step 8: Investigate the possibilities of workers’
substitution.
Step 9: Estimate the new position of workers and new
cost.
Step 10: If new cost < Pharaoh’s agent
Step 11: Replace the new agent as Pharaoh’s agent.
Where d Displacement value
g Gravity of the earth
The new position of the workers can be obtained using
the following equation.
𝑥 =
V2
0
2𝑔 sin θ (3)
The estimation of the new position can be obtained by
using the following equation
P = (Pi + d)* xЄ’i (4)
Where Pi current position
d displacement of the stone block
Є’ I normal vector that follows Levy or
normal, and uniform distribution.
IV. EFFECTS OF PERFORMANCE METRICS
IN VANET
Due to the high velocity of dynamic vehicles and
change in topology the ultimate optimum of routing
protocols is still a difficult task in VANET’s. From the
various proposed routing protocols, there is no such
protocol that is the best for all kinds of optimization
problems. Every routing algorithm is behaving
differently in sense of performance metrics. There are
various metrics like PDR, network stability, data
throughput, latency, or End to End (E2E) delay, energy
consumption, etc.
Osama Rehman et al. in [38] proposed a hybrid
scheme that improves the VANETs performance over
different traffic conditions, varying node densities, and
mobility of nodes. For the message transmission
scheme, the E2E delay reachability is improved by
10% compared to existing versions by using hybrid
relay node selection. Bukuroshe Elira et al. applied a
destination-aware context-based routing scheme
particularly in the urban environment to enhance the
PDR, varying node densities, and mobility of nodes for
message transmission scheme the E2E delay
reachability is improved by 10% compared to existing
versions by using hybrid relay node selection in [39].
In [40], Osama Rehman et al. evaluate the impact of
speed variations between vehicles and adopts two
messaging schemes like link quality and furthest
distance.
W. Kim et al. propose a routing scheme, CoRoute
for cognitive VANETs to minimize the E2E delay [41].
To establish minimum and destination they used both
the geographical location and sensed channel
information. The analysis of various routing algorithms
on VANET performance metrics was included in Table
3.
Authorized licensed use limited to: SRM University. Downloaded on August 11,2022 at 10:11:18 UTC from IEEE Xplore. Restrictions apply.

TABLE II. VARIOUS VANET ROUTING ALGORITHMS BASED ON METAHEURISTIC
Approach
Merits
Demerits
PSO-tuned AOMDV [35]
The average E2E delay dropped to 80.65%.
Network routing load dropped to 37.07%.
The packet delivery ratio dropped to 1.96%.
FF-tuned Differential Evolution [46]
Packet delivery ratio is better than AOMDV
and FF.
The average no. of hops to sink is reduced
compared to AOMDV.
WOCANET [36]
Reduced end-to-end delay in
communication.
Failed to compare with multiple algorithms.
goodness [37]
It offers alternative clustering solutions
Few parameters are missing for cluster
stability as nodes.
TABLE III. ANALYSIS OF ROUTING ALGORITHMS ON VARIOUS METRICS IN VANETS.
Approach
Authors
Simulation Parameters
Analysis
GABR [42]
Guoan Zhang et al.
Length of road segment: 2000m
Crossover operator: 0.4-0.99
Width of the road: 6m
Mutation operator : 0.0001-0.1m
Communication range: 200m
Velocity: 10-25m/s
Iterations: 100
PDR is better.
Average delay is high for a
fixed number of vehicles and
low for more number of
vehicles in the same road
segment length.
COCONUT [43]
Farhan Aadil et al.
Size of Population : 100
Maximum iterations : 150
Vehicle velocity range : 22m/s – 30m/s
Transmission range : 100m – 600m
Lane Width : 50m
Total lanes : 8
It produces 31% fewer
clusters than existing
models.
The cost of routing a packet
is minimized.
RBF neural
networks and
simulated
annealing [44]
Hosein Bagherlou et al.
Length of the road: 100km.
Number of lanes : 3
Transmission strength: 4 packets.
Radio range: 1000m.
Packet size: 1000bytes.
PDR is higher than the passing
car.
Less throughput.
Costly than pass car.
Network survival is lesser
than pass car.
Artificial swarm
algorithm [45]
Vasiliy
Krundyshev et al.
Routing protocol: AODV
Number of nodes: 20 - 100
Number of Malicious: 1 - 10
Packet Size : 1024bytes
Network traffic: CBR
Transport protocol : UDP
Simulation Time: 120sec
On Blackhole attack,
Throughput increased by
25%.
PDR is elevated by 35%.
The average delay time
decreased by 2 times almost.
On wormhole attack,
Throughput increased by
20%. PDR increased by 30%.
The average delay time
reduced by 40%.
Authorized licensed use limited to: SRM University. Downloaded on August 11,2022 at 10:11:18 UTC from IEEE Xplore. Restrictions apply.

[1] “Carnegie Mellon”. NAVslab: The Carnegie Mellon
RoboticsTheLaboratory.University NAVsigation
Institute. Retrieved 2014-12-20
[2] https://www.theverge.com/2021/2/4/22267195/ford-
electric-autonomousinvestment-29-billion
[3] https://www.cnbc.com/2021/06/16/alphabets-waymo-
raises-2point5-billion-in-new-investment-round.html
[4] https://fortune.com/2016/07/21/more-on-what-gm-paid-
to-buy-cruise-automation/
[5] https://qz.com/1599134/uber-secures much-needed-1-
billion-investment-for-self-driving-cars-unit/
[6] https://investmentmonitor.ai/automotive/hAVse -the-
wheels-come-off-investment-in-autonomous-vehicles
[7] https://en.wikipedia.org/wiki/Tesla_Autopilot
[8] Devi, S. & Malarvezhi, P. & Dayana, R. & Vadivukkarasi,
K.. (2020). A Comprehensive Survey on Autonomous
Driving Car s: A Perspective View. Wireless Personal
Communications. 114. 10.1007/s11277-020-07468-y.
[9] Fan Li, Yu Wang, “Routing in vehicular ad hoc networks:
A survey”, IEEE Vehicular Technology Magazine, Vol.2,
No.2, pp. 12-22, 2007
[10] Dhyey Patel, Mohd. Faisal Khan, Priyanka Batavia,
Sidharth Makhija, a nd M. Mani Roja, “Overview of
Routing Protocols in VANET”, International Journal of
Computer Applications, Vol.136, No.9, 2016
[11] Rao SS (2019) Engineering optimization: theory and
practice. Wiley, Hoboken
[12] Karkalos, N.E., Markopoulos, A.P., Davim, J.P. (2019).
Other Computational Methods for Optimization. In:
Computational Methods for Application in Industry 4.0.
SpringerBriefs in Applied Sciences and Technology().
Springer, Cham. https://doi.org/10.1007/978-3-319-
92393-2_4
[13] Harifi, Sasan et al. “Giza Pyramids Construction: a n
ancient- foralgorithmmetaheuristicinspired
optimization.” Evolutionary Intelligence (2021): 1-19.
[14] Neri F, C otta C, Moscato P (2011) Handbook of memetic
algorithms, vol 379. Springer, Berlin
[15] AmScialgorithms.Genetic(1992)JHHolland
267(1):66–73
[16] Geem ZW, Kim JH, Loganathan GV (2001) A new
search.harmonyalgorithm:optimizationheuristic
Simulation 76(2):60–68
[17] Das S, Suga nthan PN (2010) Differential evolution: a
survey of the state-of-the-art. IEEE Trans Evol Comput
15(1):4–31
[18] De Castro LN, Von Zuben FJ (2000) The clonal selection
algorithm with engineering a pplications. In: Proceedings
of GECCO, vol 2000, pp 36–39
[19] A. Hu ssein, H. Mostafa, M. Barrel-din, O. Sultan and A.
Khamis, "Metaheuristic optimization approach to mobile
robot path planning," 2012 International Conference on
Engineering and Technology (ICET), 2012, pp. 1-6.
[20] https://en.wikipedia.org/wiki/Tabu_search
[21] https://en.wikipedia.org/w/index.php?title=Simulated_ann
ealing&oldid=1042248468
[22] Ankit Kumar, Pankaj Dadheech, Rajani Kumari &
Vijander Singh (2019) An enhanced energy-efficient
routing protocol for VANET using a special cross over in
genetic algorithm, Journal of Statistics a nd Ma nagement
Systems, 22:7, 13491364, DOI: 10.1080/09720510.2019.
1618519
[23] Morishima K, Kuno M, Nishio A, Kitagawa N, Manabe Y,
Moto M, et al (2017) Discovery of a big void in Khufu’s
Pyramid by observation of cosmic-ray muons. Nature
552(7685):386–390
[24] Lehner M (1997) The complete pyramids. science of
Egypt’s Thames & Hudson, Lo ndon
[25] Verner M (2007) The pyramids: the mystery, culture, a nd
great monuments. Open Road+Grove/Atlantic
[26] Dawkins, R. (1976). The selfish gene, Oxford Univ.,
Oxford, U.K., 360.
[27] Geem, Zong Woo et al. “A New Heuristic Optimization
Algorithm: Harmony Search.” Simulation 76 (2001): 60 -
68.
[28] https://en.wikipedia.org/wiki/Tabu_search/algorithm
[29] swarmParticle(1995)REberhartJ,Kennedy
optimization. In: Proceedings of ICNN’95 -international
conference on neural networks, vol 4. IEEE, pp 1942–1948
[30] Dorigo M, Birattari M, Stutzle T (2006) Ant colony
optimization. IEEE Comput Intell Mag 1(4):28–39
[31] Yang XS (2009) Firefly algorithms for multimodal
optimization. In: International symposium on stochastic
algorithms. Springer, Berlin, pp 169–178 Evolutionary
Intelligence 1 3
[32] Karaboga D, Basturk B (2007) A powerful a nd efficient
algorithm for numerical function optimization: artificial
bee colony (ABC) algorithm. J Glob Optim 39(3):459–471
[33] Arif, Muhammad et al. ‘Optimization of C ommunication
in VANETs Using Fuzzy Logic and Artificial Bee
Colony’. 1 Jan. 2020: 6145 – 6157.
[34] F. Cheng and C. Shao, "Research on Artificial Fish Swarm
Clustering Algorithm in Urban Internet of Vehicles," 2020
IEEE International Conference on Smart Internet of Things
(SmartIoT), 2020, pp. 328-332, doi:
10.1109/SmartIoT49966.2020.00059.
[35] D.K. Lobiyal, C.P. Katti, A.K. G iri,Parameter Value
Optimization of Ad-hoc on Demand Multipath Distance
Vector Routing Using Particle Swarm Optimization,
Procedia Computer Science, Volume 46, 2015, Pages 151-
158, ISSN 1877-0509,
https://doi.org/10.1016/j.procs.2015.02.006 .
[36] A. Ali and S. A. A. Shah, "VANET Clustering Using Whale
Optimization Algorithm," 2019 International Symposium
on Recent Advances in Electrical Engineering (RAEE),
2019, pp. 1-5, DOI: 10.1109/RAEE.2019.8886982.
Muhammad Fahad, Far han Aadil, Zahoor-ur- Rehman,
Salabat Khan, Peer Azmat Shah, Khan Muhammad, Jaime
Lloret, Haoxiang Wang, Jong Weon Lee, Irfan Mehmood,
Grey wolf optimization based clustering algorithm for
vehicular ad-hoc networks, Computers & Electrical
Engineering, Volume 70, 2018, Pages 853-870, ISSN 0045-
7906, https://doi.org/10.1016/j.compeleceng.2018.01.002.
[37] Osama Rehman, Mohamed Ould-Khaoua, A hybrid relay
node selection scheme for message dissemination in
VANETs, Future G eneration Computer Systems, Volume
93, 2019, Pages 1-17, ISSN 0167-739X,
https://doi.org/10.1016/j.future.2018.10.042 .
[38] Bukuroshe Elira, K.P. Keerthana, K. Balaji, Clustering
scheme and destination aware context-based routing
protocol for VANET, International Journal of Intelligent
Networks, Volume 2, 2021, Pages 148-155, ISSN 2666-
6030.https://doi.org/10.1016/j.ijin.2021.09.006.(https://ww
w.sciencedirect.com/science/article/pii/S26666030210002
21).
[39] Rehman, O., Qureshi, R., Ould-Khaoua, M., & Niazi, M.
F. (2020). Analysis of mobility speed impa ct on end-to-
end communication performance in VANETs. Vehicular
Communications, 26,[100278]. https://doi.org/10.1016/j.v
ehcom.2020.100278.
[40] W. Kim, M. Gerla, S.Y. Oh, K. Lee, A. Kassler, Coroute:
a new cognitive any path vehicular routing protocol,
V. CONCLUSION
VANETs
can
make
our
lives
more
secure
by
adding
more
vehicles
and
making
driving
safe.
Routing
plays
a
very
crucial
role
in
VANET
communication.
Designing an optimization routing
algorithm for VANET applications is not easy.
This
paper
includes
a
survey
on
various
existing
metaheuristic-based routing
algorithms
and
includes
various
algorithms’
behavior
under
circumstances
like
E2E
delay,
PDR,
throughput,
and
cluster
stability.
This
paper
also
includes
a
review
of
various algorithms applied in the area of VANET.
REFERENCES
Authorized licensed use limited to: SRM University. Downloaded on August 11,2022 at 10:11:18 UTC from IEEE Xplore. Restrictions apply.

Wireless Communication, Mobile Computing 11 (12)
(2011) 1588–1602.
[41] XinmingDuan,WeiWu,MinZhang,Guoan
Huang, PerceptionBasedAlgorithm"Genetic QoS
VANETs",forProtocolRouting Wireless
MobileandCommunications
Computing, vol. 2018, Article
ID 3897857, 10 pages, 2018. https://doi.org/10.1155/2018
/3897857
[42] Aadil F, Bajwa KB, Khan S, Chaudary NM, Akram A
(2016) CACONET: Ant Colony Optimization (ACO)
Based Clu stering Algorithm for VANET. PLoS ONE
11(5): e0154080. doi:10.1371/journal. pone.0154080
[43] Bagherlou, H., Ghaffari, A. A routing protocol for
vehicular ad hoc networks using simulated annealing
algorithm and neural networks. J Supercomput 74, 2528–
2552 (2018). https://doi.org/10.1007/s11227-018-2283-z
[44] V. Krundyshev, M. Kalinin and P. Zegzhda, "Artificial
swarm algorithm for VANET protection against routing
attacks," 2018 IEEE Industrial Cyber-Physical Systems
795pp.2018,(ICPS), -800, DOI:
10.1109/ICPHYS.2018.8390808.
[45] https://doi:10.35940/ijrte.D1053.1284S519
[46] Sarubbi, Joao FM, Flavio VC Martins, a nd Cristiano M.
Silva. "A genetic algorithm for deploying roadside units in
VANETs." EvolutionaryonCongressIEEE2016
Computation (CEC). IEEE, 2016.
[47] XinmingDuan,WeiWu,MinZhang,Guoan
Huang, PerceptionQoSAlgorithm"Genetic Based
VANETs",forProtocolRouting Wireless
C Mobileandommunications
Computing, vol. 2018, Article
ID 3897857, 10 pages, 2018. https://doi.org/10.1155/2018
/3897857
[48] Faraj, Marcelo F., et al. "A Memetic Algorithm Approach
to Deploy RSU s Based on the Gamma Deployment
Metric." EvolutionaryonCongressIEEE2018
Computation (CEC). IEEE, 2018.
[49] Ravie Chandran Muniyandi, Mohammad Kamrul Hasan,
Mustafa Raad Hammoodi, Ali Maroosi, "An Improved
for Proactive RoutingSearch AlgorithmHarmony
VANET",inProtocol AdvancedofJournal
Transportation, vol. 2021, Article
ID 6641857, 17 pages, 2021. https://doi.org/10.1155/2021
/6641857
[50] Er. Jayant Vasu, Dr. Gaurav Tejpal and Dr. Sonal Sharma,
Adaptive Differential Evolution Based Attack Detection
Protocols for Vanets, International Journal of Mechanical
Engineering and Technology 8(10), 2017, pp. 217–228.
http://www.iaeme.com/IJMET/issues.asp?JType=IJMET
&VType=8&IType=10
[51] J. Hu, C. Chen, T. Qiu, M. Atiquzzaman and Q. Pei,
"Elastic and Inelastic Content Distribution Based on
Clonal Selection in VANETs," 2019 IEEE Global
Communications Conference (GLOBECOM), 2019, pp. 1-
6, DOI: 10.1109/GLOBECOM38437.2019.9013918.
[52] Moridi, Elham, a nd Hamid Barati. "RMRPTS: a reliable
multi- protocolroutinglevel insearchtabuwith
VANET." Telecommunication Systems 65.1 (2017): 127-
137.
[53] Bagherlou, H., Ghaffari, A. A routing protocol for
vehicular ad hoc networks using simulated annealing
algorithm and neural networks. J Supercomput 74, 2528–
2552 (2018). https://doi.org/10.1007/s11227-018-2283-z
Authorized licensed use limited to: SRM University. Downloaded on August 11,2022 at 10:11:18 UTC from IEEE Xplore. Restrictions apply.