JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 38, 41-56 (2022)

DOI: 10.6688/JISE.202201_38(1).0003

PSO-DQ: An Improved Routing Protocol Based

on PSO using Dynamic Queue Mechanism for MANETs

PRIYANKA KUMARI1 AND SUDIP KUMAR SAHANA2

Department of Computer Science and Engineering
Birla Institute of Technology
Jharkhand, 835215 India
E-mail: phdcs10004.16@bitmesra.ac.in1; sudipsahana@bitmesra.ac.in2

The growth of mobile device technologies has given rise to widespread applications
that led us to economic wireless networks, including with and without infrastructure. Effi-
cient routing with Quality-of-Service (QoS) constraints is a challenging issue in substantial
infrastructure-less and dynamic networks. To improve QoS constraints for such a network
is an NP-complete problem. It is observed that Particle Swarm Optimization (PSO) is one
of the most potent swarm-based optimization techniques to solve NP problems. Hence,
PSO is chosen to boost QoS constraints and provide more reliable routes than existing on-
demand routing protocols. This paper has proposed a PSO-based routing facility that uses
a dynamic queue mechanism for efficient routing considering enriched QoS constraints.
The uniqueness of the proposed technique is selecting the fitness function that is dynamic
in nature and determined based on the data obtained by the successor node. The queue size
is maintained dynamically to minimize the data drop. The simulation results revealed that
the proposed algorithm performs better than the existing conventional algorithms like Ad-
hoc On-Demand Distance Vector (AODV), Destination-Sequenced Distance Vector Rout-
ing (DSDV), and metaheuristics like ACO, PSO, QoRA, Enhanced-Ant-AODV, and
Cuckoo Search Optimization AODV (CSO-AODV) in terms of packets sent, packets re-
ceived, PDR, end-to-end delay and routing overhead.

Keywords: routing, dynamic queue, MANETs, ACO, PSO, quality of service

1. INTRODUCTION

A mobile ad-hoc network (MANET) is a self-configuring and framework-less network

functioning without centralized administration where nodes move randomly. Conse-
quently, the topology of the network may experience speedy and random changes [1]. As
nodes in MANET usually have bounded communication ranges, some nodes cannot com-
municate independently. Data packets are sent from the source node to the destination node
to support the multiple intermediate nodes in the communication network. MANETs in-
clude indispensable multiple hops. Each node in MANETs is responsible for acting as a
host to generate data packets and a router to forward those data packets [2]. Each node has
its wireless interface to communicate with the other. Fig. 1 exemplifies that nodes A and
C are not within the range with each other, and thus node B can be used to forward packets
between node A and node C. Here, node A and node C determine the route through node
B; node B will act as a router. Mobile ad-hoc network allows its users to use it more fre-
quently in those circumstances where a static framework is nonexistent or a static frame-
work is challenging to build, like Military Environments, Personal Area Networking, Ci-
vilian Environments, or Emergency Operations.
Received August 27, 2020; revised February 15, 2021; accepted March 23, 2021.
Communicated by Xiaohong Jiang.

41
42 PRIYANKA KUMARI AND SUDIP KUMAR SAHANA

Fig. 1. Example of mobile ad hoc network.

But due to autonomous infrastructure-less limited resource availability, lack of cen-

tralized administration, and dynamic nature of MANETs, routing and providing QoS re-
quirements becomes a tedious task in communication networks [3-6]. Routing is selecting
a route between the source nodes to the destination nodes through which the source node
delivers the data packets to the destination nodes in the network [7]. The traditional and
well-known algorithms such as the Bellman-Ford dynamic programming algorithm and
Dijkstra greedy algorithm has been used to find the shortest route, but they have the fol-
lowing significant drawbacks:

(i) They are not appropriate for the networks having negative weights of edges.
(ii) They are not fit for large and dynamic network topologies.

With the rapid development of network services in modern society, routing has be-
come more and more popular. Therefore, selecting the appropriate routing protocol is an
important and tedious task that must satisfy QoS parameters which are also known as per-
formance parameters of a network like throughput, delay, jitter, packet delivery ratio
(PDR), routing overhead (RO), and reliability [8]. QoS constraints directly affect routing;
that’s why routing is such a critical facet of network communication. The objective of
selecting a routing protocol is to find the optimal routes with maximum PDR, minimum
end-to-end delay, and minimum RO. When the source node sends the data packets within
the network, QoS constraints may affect the network’s performance. So, some extra means
are required to tackle such situations in highly dynamic and extensive MANETs. Many
routing protocols have been developed like AODV [9], DSR [10], OLSR [11], and DSDV
[12]. However, in the case of an extensive and dynamic network and a combination of two
or more QoS constraints, either additive, multiplicative, or a mixture of additive or multi-
plicative metrics contribute to NP-complete in nature and are not found suitable.
Most of the research is targeted towards considering either a single QoS metric or two
QoS metrics, and very few are considering three metrics. However, real-time communica-
tions require a minimum end-to-end delay, maximum available bandwidth, high PDR, and
low RO. A possible solution to these kinds of difficulties, which cannot be resolved with
classical methods, is the application of stochastic optimization techniques. Stochastic op-
timization techniques are usually categorized into evolutionary algorithms, like genetic
algorithm (GA) and nature-inspired algorithms like ant colony optimization (ACO) and
particle swarm optimization (PSO). Compared to GA and ACO, the advantages of PSO
are faster, cheaper, easy to implement, and a smaller number of parameters to be adjusted
[13].
 PSO-DQ 43

PSO, the population-based stochastic search algorithm, was first introduced by Dr

Kennedy and Dr Eberhart in 1995 and based on the natural behaviour of birds flocking. In
PSO, each particle flies through the multidimensional space, adjusts its position in every
step based on its own and that of its peers’ experiences until the entire swarm seeks an
optimal solution.
Here, a dynamic queue-based PSO optimization is proposed to improve the QoS pa-
rameters. It is essential to understand routing protocols followed by a summary of a few
already existing schemes listed in Section 2. The proposed dynamic queue-based PSO is
presented in Section 3. The simulation outcomes of the proposed scheme are compared
with AODV, DSDV, ACO, PSO, QoRA, Enhanced-Ant-AODV, and CSO-AODV meth-
ods in Section 4.

2. RELATED WORK

Researchers have offered numerous routing techniques to meet the dynamic topology
challenges, which finds the optimal path with improving QoS constraints [14]. These pro-
tocols can be categorized into three types: − Conventional methods [15], evolutionary
methods [16], and swarm intelligence-based methods [17].
A brief overview of the preceding works considering one or two QoS constraints for
searching the optimal route in MANET is presented in this section. We have given an
overview of existing conventional, evolutionary, and swarm-based routing protocols in this
section.

2.1 Conventional Routing Protocols

Conventional routing protocols are categorized into three classes: −proactive, reactive,
and hybrid routing protocols. Every node has its routing tables in the proactive routing
protocols to contain routing information to every other node. It is periodically updated
when a node observes any significant network topology change. However, the reactive
routing protocols find routes only on-demand of the network. At the same time, hybrid
routing protocols use the best features of both routing protocols.
DSDV [12] routing protocol based on Bellman-Ford routing algorithm [18] with spe-
cific modification such as loop-free, it offers the shortest single path to the final node.
There occurs a large amount of routing overhead to the network if there are many nodes or
extensively mobile, and it also consumes a large amount of bandwidth to update the routing
data at each node. DSR [10] protocol is a combination of route discovery and route mainte-
nance mechanism of the network route. In DSR, the network’s overhead routing increases
with increasing node density because each packet carries the full address of the whole route.
Due to this reason, DSR is not appropriate for bulky and highly dynamic networks. AODV
[9, 19, 20] is a combination of DSDV and DSR algorithm. It adopts the sequence number
procedure of DSDV and the route discovery and route maintenance mechanism of DSR. It
is adaptable for highly dynamic topologies, but it consumes more bandwidth and intro-
duces additional delays when the network size increases. Zone Routing Protocol (ZRP)
[21] is a hybrid routing protocol. It behaves like a table-driven routing protocol for the
routing zone's enormous value and an on-demand routing protocol for the small value of
the routing zone as it reduces routing overhead. OLSR [11] table-driven routing protocol,
44 PRIYANKA KUMARI AND SUDIP KUMAR SAHANA

whose key idea is Multipoint Relays (MPRs) to reduces control overhead. As per the au-
thors, in this algorithm, the shortest paths to every destination are available without any
delay when data transmission is required. Here, we have analyzed that the conventional
routing protocols are not sufficient in large and dynamic constraints. These are designed
without explicit consideration in quality-of-service for the generated routes. Table 1 pro-
vides a sequential summary of the essential characteristics of conventional routing proto-
cols.

Table 1. Chronological summary of essential characteristics of conventional routing protocols.

Existing Routing Multiple
Advantages Disadvantages
Protocols Approach Routes
DSVD  High overhead
1994 Proactive No  Single shortest path  Consumes more bandwidth
[12]
 Multiple Routes  Scalability problems
DSR  Promiscuous over-  Flooding
1996 Reactive Yes hearing  Large delays
[10]  Save a considerable
amount of bandwidth
 Suitable to highly  Large delays
AODV dynamic topologies  Not suitable for large net-
1999 Reactive No  Less routing over- works
[9] head than DSR  Consumes more bandwidth
 Reduces the over-  Behaves like proactive rout-
ZRP
head ing protocol for a large network
2000 Hybrid No
[21]  Behaves like reactive routing
protocol for a small network
 Reduced control  Reduces the control overhead
OLSR overhead  It needs more time to re-dis-
2001 Proactive No covering the broken link
[11]  Requires more processing
power

2.2 Evolutionary and Swarm Based Routing Protocols

The genetic algorithm is a speculative search technique and an evolutionary approach

encouraged by the Darwinian principles [16] of natural selection genetics, which has
shown several characteristics particularly useful for routing search in MANETs [22, 23].
GA, with its growing nature, optimizes the shortest path problem by producing improved
results with the given solutions. Barolli et al. [24] proposed a genetic algorithm (GA) based
routing method for MANETs (GAMAN) to use end-to-end delay and packet dropping rate
as QoS constraints. Sanghoun Oh et al. [25] proposed a Genetic-Inspired Multicast Rout-
ing Optimization Algorithm, which increases the network efficiency only in terms of band-
width and delay constraints.
Swarm intelligence is a computational intelligence technique that contains the com-
bined performance of self-directed agents that nearby communicate in a scattered atmo-
 PSO-DQ 45

sphere to solve a specified problem in the expectation for asset solution to the problem.
Recently, researchers show their concern in using swarm intelligence (SI) [26, 27] for rout-
ing in MANET. Ant colony optimization (ACO), particle swarm optimization (PSO), bac-
terial foraging optimization (BFO), and artificial bee colony (ABC) are examples of swarm
intelligence. ARA [28] ant-based routing algorithm has proposed to reduce overhead in
routing. AntHocNet [29] is a hybrid routing protocol that uses reactive routing protocols
for path set and proactive routing protocols to maintain the path. HOPNET [30] is an ant-
based routing algorithm that borrows the features of DSR and ZRP and gives better per-
formance than AODV, ZRP, AntHocNet. AMQR [31] reactive routing algorithm based on
ACO for ad hoc networks has proposed extending the path with high preference probability
for the minimum delay, maximum bandwidth, and minimum hop count. AMAR [32] uses
the combining ideas of artificial intelligence (AI) and multipath routing, in which the al-
gorithm for improving network performance is achieved. Singh et al. [33] have provided
a comparative analysis for ACO-based algorithms in MANETs for various QoS metrics.
Hemlata, Uradea and Patel [34] concluded that the dynamic PSO gives better-optimized
value to multi-objective optimization problems. QoRA [35] reactive routing protocol based
on local SNMP (Simple Network Management Protocol) to find the path satisfies QoS
constraints by finding multiple ways. CSO-AODV [36] routing protocol based on enhanc-
ing Cuckoo Search (CS) technique gives better result in terms of PDR, packet drops, and
overhead. Enhanced-Ant-AODV [37] uses combining AODV and ACO ideas to improve
QoS constraints and provide better results than AODV, DSR, and Enhanced-DSR in terms
of PDR, throughput, and delay.
Hybrid PSO-GA [13] multicast routing algorithm combines PSO and GA’s strengths
to balance natural selection and good knowledge sharing to provide a robust and efficient
search of the solution space. Patel et al. [38] proposed a multicast routing optimization
based on ACO and PSO, which utilize the collective and coordination process for mobile
agents attached to each pattern to satisfy the QoS constraints. Table 2 provides a sequential
summary of the essential characteristics of the swarm and evolutionary-based routing pro-
tocols.

Table 2. Chronological summary of essential characteristics of evolutionary and swarm

based routing protocols.
Existing
Routing Multiple Compared
Schemes/ Advantages Disadvantages
approach Routes with
Protocols
AODV,  Less overhead  Not support high mo-
ARA
Reactive Yes DSR, bility
2002 [28]
DSDV
 End-to-end delay  Supports only for
GAMAN
Reactive Yes −  Packet dropping small and medium net-
2003 [24]
rate. works.
AntHocNet AODV  Improve PDR,  More overhead
Hybrid Yes
2005 [29] delay, jitter
Chen’s algo-  Satisfies QoS  Not mentioned about
GA Routing
Reactive Yes rithm (bandwidth, end to PDR
2006 [25]
end delay)
46 PRIYANKA KUMARI AND SUDIP KUMAR SAHANA

Table 2. (Cont’d) Chronological summary of essential characteristics of evolutionary and

swarm based routing protocols.
Existing
Routing Multiple Compared
Schemes/ Advantages Disadvantages
approach Routes with
Protocols
 High communication
AODV,
HOPNET  Improve PDR and complexity
Hybrid Yes ZRP, Ant-
2009 [30]
HocNet
delay  Not fit for the large
network
 Provides good PDR
 Reduces delay
AMQR AODV,  High overhead
Reactive Yes  Reduces jitter
2011 [31] AntHocNet  Congestion problem
 Supports node
mobility
PSO-GA PSO,  Improve PDR
Hybrid Yes
and delay  High overhead
2011 [13] GA
AODV  Improve PDR and
AMAR
Hybrid Yes OLSR delay  High overhead
2012 [32]
AntHocNet
 Satisfies delay
HACOPSO PSOTREE, and delay jitter  Not mentioned about
Reactive Yes
2014 [38] TGBACA PDR

 Avoid network
QoRA congestion  Jitter increases as net-
Reactive Yes AODV  Avoid packet
2015 [35] work size increases.
loss
AODV,  Supports scala-
CSO-AODV  Not mentioned about
Reactive Yes ACO, bility and mobility
2016 [36] the delay.
PSO
Enhanced- AODV, DSR,  Improve PDR,  Not mentioned about
Ant-AODV Reactive Yes Enhanced- throughput and de- packet loss ratio and
2018 [37] Ant-DSR lay jitter.

3. PROPOSED METHODOLOGY

Particle swarm optimization is an efficient method to find a reliable route while the
network is dynamic. In this proposal, PSO is used to solve QoS’s network constraints be-
cause it provides a more reliable route than existing on-demand routing protocols. Here,
PSO operates based on a dynamic fitness function by calculating successful data received
by the successor nodes from the predecessor nodes. In this proposed approach, the dynamic
queue is used to improve the network QoS parameters. In the dynamic queue method, de-
mand-based intermediate nodes update their queue size to minimize data dropping from
the network. It also plays a critical role in memory management. If the bandwidth of all
links are equal, then the queue size is less reserved. On the other hand, if the higher varia-
tion of bandwidth between link and data rate is consistent, then the queue is highly needed.
Thus, this approach makes the dynamic queue mechanism beneficial to maintain the net-
work QoS constraints. As the proposed model’s formal notion is framed, the fitness value
 PSO-DQ 47

calculation for PSO is based on the Eq. (1) to pick the particular path using the dynamic
queue method to review the network’s dropping tail.

i=1( fwdi − fsec r _ i )

n

f f _k = 1− Rec prd
(1)

In PSO, all particles are initiated randomly. Suppose xit denotes the position vector of
particle i at time t. Each particle adjusts its position in the multidimensional search space
(xmin, xmax) according to Eq. (2) based on its own experience and its neighbours’ experience.
All particles are evaluated to compute the pbest (best value of each particle) and gbest (best
value of particle in the entire swarm) according to Eqs. (3) and (4) respectively. The ve-
locity of particle i is updated according to Eq. (5). After updating the position and velocity
of the particle according to Eqs. (2) and (5) respectively, evaluate the fitness function ff_k
according to Eq. (1). Various symbols and their meaning are summarized in Table 3. Fig.
2 shows the schematic diagram of the proposed PSO-DQ approach.

xit +1 = xit + vit +1 with xi0  U ( xmin , xmax ) (2)

t +1
t +1 pbest if f ( xit +1 )  pbest
t

.i =
pbest .i .i (3)
xit +1 if f ( xit +1 )  pbest
t
.i

Gbest = min{ pbest

t
.i } where i [0,1,..., n]& n  1
(4)

vijt +1 = vijt + C1r1tj [ pbest

t
,i − xij ] + C2 r2 j [Gbest − xij ]
t t t t
(5)

Table 3. Symbols used and their meaning.

ff_k fitness function to find the best location
fwdi forward to next neighbor by ith node
fsecr_i failure in receiving the data by successor node
Recprd Successful data received from a predecessor
xit The position vector of particle i at time t
vit The velocity vector of particle i drives the optimization process and reflects
own experience knowledge and social experience knowledge from all parti-
cles.
U(xmin, xmax) Uniform distribution where is its minimum xmin and xmax maximum values, re-
spectively.
vtij The velocity vector of a particle in dimension j at time t
xtij The position vector of particle i in dimension j at time t
ptbest,i Personal best position of particle i in dimension j found from initialization
through time t
Gtbest Global best position of the particle in dimension j found from initialization
through time t
C1 and C2 Acceleration constants
rt1j and rt2j Random numbers from uniform distribution U(0, 1) at time t
48 PRIYANKA KUMARI AND SUDIP KUMAR SAHANA

Fig. 2. Schematic diagram of the proposed PSO-DQ approach.

 PSO-DQ 49

The pseudo-code of the proposed PSO-DQ algorithm is described as follows:

Proposed PSO-DQ for QoS under MANET

Input: Zn: n number of mobile nodes
qi: source nodeZn
ri: receiver nodeZn
ik: kth intermediate nodeZn
dn: queueZn
: Radio Range 550m2 Znk-1 = Zk-1
Rp: PSO
ff: fitness function to find best location
pbesti: past best location (initial = 0.0)
Rv: random position
Sp: search space 550m2
li: current location
Recprd: success full receiver data from predecessor
fwdi: forward to next neighbor by ith node
fsecr_i: failure to receive by successor node
Par[]: population array
Vi: velocity of ith node
pathi: path between qi to ri
Output: Data Send, Data Receives, nrl, delay, pdr
Procedure:
Step1: Zn initialize Par[], Rv under Sp
qi execute route_req(qi, ri, Rp)
Step2: for i = 1 to n
If ik in  && ik! = ri then
i=1( fwdi − f secr_i )
n

Calculate f f _ k = 1 − Rec prd

If ff_k > pbestk then

pbestk = ff_k
ik forward (qi, ri, Rp) to next-neighbor
lnew_k  lold_k + Vkd
Else
pbestk
li  lold_k
Discard route packet
End if
Else if ik in  && ik == ri then
ri store (pathi, ff_i, lipathi)
ri generate reverse path to qi
qi receives ack from ri
Send-data (qi, ri, pathi)
Else
ri not reachable or not in range
End if
50 PRIYANKA KUMARI AND SUDIP KUMAR SAHANA

End for
Step3: send-data (qi, ri, data)
Step4: qi check dnj=1 those node in pathi
Step5: for j 1 to n
If data receives at ith node&& ij! = ri then
dj  dj + 1
ij check route table to send next-successor
Else if data forward from ith node&& ij! = ri then
dj  dj − 1
ij forward (qi, ri, data) to next-successor
Else if data receives at ith node&& ij == ri then
di  di + 1
Retrieve data from di
Send ack to qi from reverse pathi
Else
ri not in  or pathi break
Connection terminate
End if
End for

We used the event-driven network simulator NS2 version 2.31 to evaluate the effi-
ciency of the results obtained. The simulation area is 10001000 square meters with node
numbers 50, 75, and 100, where the nodes are placed randomly. Table 4 shows the other
network simulation parameters.

Table 4. Parameters for the simulation scenario.

Parameters Used in Simulation
Network type MANET
Simulator Ns2
MAC type IEEE 802.11b
Area 1000*1000 sq. m
Routing protocol AODV, DSDV, ACO, PSO, QoRA, Enhanced-
Ant-AODV, CSO-AODV, PSO-DQ
No. of nodes 50, 75, 100
Transport layer TCP, UDP
Application layer FTP, CBR
Packet size 512 Bytes
No. of connection 15-20
Antenna Model OmniAntenne
Propagation model TwoRayGround
Queue mechanism Droptail/PriQueue
Mobility Model Random Waypoint
Simulation time 480 sec
Pause time 10 sec
 PSO-DQ 51

Performance measuring QoS parameters:

(1) Data sending: Data sending depends on the data rate and path availability between the
sources to the receiver nodes.
(2) Data received: The number of data received depends on network behaviour such as
path availability, queue utilization, bandwidth, etc.
(3) PDR: PDR is a ratio of total data packets received successfully to the total data packets
sent from the initial node to the final node.

Total received packets

PDR = 100 (6)
Total sent packets

(4) End-to-end delay: It is the average time of the data packet to transmit successfully
from the initial node to the final node.
(5) Routing overhead (RO): It is the ratio of the total packet sent to the number of control
packets sent.

No. of sent packets

RO = (7)
No. of control packets sent

4. RESULTS AND DISCUSSIONS

(1) Data Sending Analysis: Fig. 3 shows the performance comparison of data sending of
DSDV, AODV, ACO, PSO, QoRA, Enhanced Ant-AODV, CSO-AODV, and PSO-DQ. In
this graph, PSO-DQ gives better results concerning node variation as compared to all the
existing protocols.
No of Packets

No of Packets

Fig. 3. No. of nodes vs. data sending analysis. Fig. 4. No. of nodes vs. data receiving an-
alysis.

(2) Data Receiving Analysis: The number of data received depends on network behavi-
our such as path availability, queue utilization, bandwidth, etc. Fig. 4 shows the perfor-
mance comparison of data received by DSDV, AODV, ACO, PSO, QoRA, Enhanced Ant-
AODV, CSO-AODV, and PSO-DQ. The performance comparison shows that the dynamic
queue with the PSO-based mechanism gives excellent results compared to the existing
routing protocol.
 52 PRIYANKA KUMARI AND SUDIP KUMAR SAHANA

(3) PDR Performance Analysis: PDR is a significant factor in analyzing network be-
haviour because it measures the percentage of data received at the receiver end. PDR not
only depends on the bandwidth availability, but it also depends upon the network con-
gestion, the number of route changes, buffering at the intermediate nodes. If network
congestion is higher or routes frequently change, or buffer is full, packet delivery perfor-
mance goes down. Fig. 5 shows PDR performance in three scenarios at node numbers 50,
75, and 100. Through the graph, we conclude that PSO-DQ slightly improves the packet
delivery ratio as compared to DSDV, AODV, ACO, PSO, QoRA, Enhanced Ant-AODV,
CSO-AODV method as well as they are found to be having a positive influence on other
network parameters.
(4) Delay Performance Analysis: The delay in the network depends on the communi-
cation link, queuing process of data, channel availability, retransmission of data packets
and link break, or other reasons. In MANET, nodes change their location every second and
exchange information with each other, resulting from dynamic delays per-packet. Fig. 6
shows the performance comparison of the end-to-end delay of DSDV, AODV, ACO, PSO,
QoRA, Enhanced Ant-AODV, CSO-AODV, and PSO-DQ. This graph shows the average
delay in the duration ranging in a millisecond and indicates variation based on network
size. We conclude that as the network size increases, the average network delay of pro-
posed PSO-DQ is lower as compared to the existing protocols.
(5) Overhead Analysis: The Overhead is directly proportional to delay in the network,
and it occurs due to frequent node motion, network congestion, or heavy traffic. The
overhead in the network increases due to frequent link breakage in the network. Fig. 7
shows the performance comparison based on the overhead of DSDV, AODV, ACO, PSO,

Packet Delivery Radio Analysis Average Delay in ms Analysis

Percentage

Delay(ms)

Fig. 5. No. of nodes vs. PDR. Fig. 6. No. of nodes vs. delay (ms).
overhead

Fig. 7. No. of nodes vs. Routing overhead.

 PSO-DQ 53

QoRA, Enhanced Ant-AODV, CSO-AODV, and PSO-DQ. Though there is a little hike in
the PSO-DQ, the efficiency of packet receiving has improvised. The promenade of
overhead is occurring due to the routing balancing of the load in the network. From the
graph, we conclude that as the network size increases, the routing overhead of PSO-DQ is
comparatively less than DSDV, AODV, ACO, PSO, QoRA Enhanced-Ant-AODV.

5. CONCLUSIONS

The novelty of this work is the use of swarm-based powerful optimizer PSO along
with changing queue mechanism for minimizing packet drop and improving QoS con-
straints such as packet sending, packet receiving, PDR, delay, and routing overhead. Im-
proving QoS is much significant and desirable aspect of MANETs. Nodes with a fixed
queue length might cause the possibility of higher dropping, so the proposed routing
schemes are planned based on varying queue size. The proposed scheme increases network
efficiency, as well. The improvements in this protocol are evaluated by the network’s QoS
performance metrics and compared with AODV, DSDV, ACO, PSO, QoRA, Enhanced-
Ant-AODV, and CSO-AODV routing algorithms. The results obtained certify the suprem-
acy of the proposed dynamic queue-based PSO over compared algorithms.

REFERENCES

1. A. A. A. Radwan, T. M. Mahmoud, and E. H. Houssein, “Evaluation comparison of

some ad hoc networks routing protocols,” Egyptian Informatics Journal, Vol. 12, 2011,
pp. 95-106.
2. I. Chlamtac, M. Conti, and J. J. N. Liu, “Mobile ad hoc networking: Imperatives and
challenges,” Ad Hoc Networks, Vol. 1, 2003, pp. 13-64.
3. S. Chakrabarti and A. Mishra, “QoS issues in ad hoc wireless networks,” IEEE Com-
munications Magazine, Vol. 39, 2001, pp. 142-148.
4. B. Zhang and H. T. Mouftah, “Qos routing for wireless ad hoc networks: problems,
algorithms, and protocols,” IEEE Communications Magazine, Vol. 43, 2005, pp. 110-
117.
5. Y. Xu, J. Liu, Y. Shen, J. Liu, X. Jiang, and T. Taleb, “Incentive jamming-based secure
routing in decentralized Internet of Things,” IEEE Internet of Things Journal, Vol. 8,
2021, pp. 3000-3013.
6. Y. Xu, J. Liu, Y. Shen, X. Jiang, Y. Ji, and N. Shiratori, “QoS-aware secure routing
design for wireless networks with selfish jammers,” IEEE Transactions on Wireless
Communications, Vol. 20, 2021, pp. 4902-4916.
7. E. M. Royer and C.-K. Toh, “A review of current routing protocols for ad hoc mobile
wireless networks,” IEEE Personal Communications, Vol. 6, 1999, pp. 46-55.
8. S. Aswale and V. R. Ghorpade, “Survey of QoS routing protocols in wireless multi-
media sensor networks,” Journal of Computer Networks and Communications, Vol.
2015, 2015, No. 824619.
9. C. E. Perkins and E. M. Royer, “Ad-hoc on-demand distance vector routing,” in Pro-
ceedings of the 2nd IEEE Workshop on Mobile Computer Systems and Applications,
54 PRIYANKA KUMARI AND SUDIP KUMAR SAHANA

1999, pp. 90-100.

10. D. B. Johnson and D. A. Maltz, “DSR: The dynamic source routing protocol for multi-
hop wireless ad hoc networks,” in Ad Hoc Networking, C. E. Perkins, ed., Chapter 5,
Addison-Wesley, 2001, pp. 139-172.
11. P. Jacquet, P. Mühlethaler, T. H. Clausen, et al., “Optimized link state routing protocol
for ad hoc networks,” in Proceedings of IEEE International Multi-Topic Conference,
2001, pp. 62-68.
12. C. E. Perkins and P. Bhagwat, “Highly dynamic destination-sequenced distance-vec-
tor routing (dsdv) for mobile computers,” in Proceedings of ACM SIGCOMM Com-
puter Communication Review, Vol. 24, 1994, pp. 234-244.
13. R. F. Abdel-Kader, “Hybrid discrete PSO with GA operators for efficient QoS-mul-
ticast routing,” Ain Shams Engineering Journal, Vol. 2, 2011, pp. 21-31.
14. P. Kumari and S. K. Sahana, “An efficient swarm-based multicast routing technique
− Review,” in Proceedings of the International Conference on Computational Intelli-
gence in Data Mining, 2019, pp. 123-134.
15. M. Alam, A. H. Khan, and I. R. Khan, “Swarm intelligence in MANETs: a sur-
vey,” International Journal of Emerging Research in Management and Technology,
Vol. 5, 2016, pp. 141-150.
16. Y. S. Yen, Y.-K. Chan, H.-C. Chao, and J. H. Park, “A genetic algorithm for energy-
efficient based multicast routing on MANETs,” Computer Communications, Vol. 31,
2008, pp. 2632-2641.
17. M. Saleem, G. A. Di Caro, and M. Farooq, “Swarm intelligence based routing protocol
for wireless sensor networks: Survey and future directions,” Information Sciences,
Vol. 181, 2011, pp. 4597-4624.
18. R. Bellman, “Dynamic programming and stochastic control processes,” Information
and Control, Vol. 1, 1958, pp. 228-239.
19. E. M. Belding-Royer and C. E. Perkins, “Evolution and future directions of the ad hoc
on-demand distance-vector routing protocol,” Ad Hoc Networks, Vol. 1, 2003, pp.
125-150.
20. D. B. Johnson, “Routing in ad hoc networks of mobile hosts,” in Proceedings of IEEE
1st Workshop on Mobile Computing Systems and Applications, 1994, pp. 158-163.
21. J. Lundberg, “Routing security in ad hoc networks,” Helsinki University of Technol-
ogy, http://citeseer.nj.nec.com/400961.html, 2000.
22. L. Li and C. Li, “Genetic algorithm-based QoS multicast routing for uncertainty in
network parameters,” in Proceedings of Asia-Pacific Web Conference, 2003, pp. 430-
441.
23. R.-H. Hwang, W.-Y. Do, and S.-C. Yang, “Multicast routing based on genetic algo-
rithms,” Journal of Information Science and Engineering, Vol. 16, 2000, pp. 885-901.
24. L. Barolli, A. Koyama, and N. Shiratori, “QoS routing method for ad-hoc networks
based on genetic algorithm,” in Proceedings of IEEE 14th International Workshop on
Database and Expert Systems Applications, 2003, pp. 175-179.
25. S. Oh, C. Ahn, and R. S. Ramakrishna, “A genetic-inspired multicast routing optimi-
zation algorithm with bandwidth and end-to-end delay constraints,” in Proceedings
of International Conference on Neural Information Processing, 2006, pp. 807-816.
26. E. Bonabeau, M. Dorigo, and G. Theraulaz, Swarm Intelligence: from Natural to Ar-
tificial Systems, Oxford University Press, UK, 1999.
 PSO-DQ 55

27. P. Kumari and S. K. Sahana, “QoS-based ACO routing protocols in MANETs: A

review,” in Proceedings of the 4th International Conference on Microelectronics
Computing and Communication Systems, 2021, pp. 329-340.
28. M. Gunes, U. Sorges, and I. Bouazizi, “ARA-the ant-colony based routing algorithm
for MANETs,” in Proceedings of IEEE International Conference on Parallel Pro-
cessing Workshop, 2002, pp. 79-85.
29. G. D. Caro, F. Ducatelle, and L. M. Gambardella, “AntHocNet: an adaptive nature‐
inspired algorithm for routing in mobile ad hoc networks,” European Transactions on
Telecommunications, Vol. 16, 2005, pp. 443-455.
30. J. Wang, E. Osagie, P. Thulasiraman, and R. K. Thulasiram, “HOPNET: A hybrid ant
colony optimization routing algorithm for mobile ad hoc network,” Ad Hoc Networks,
Vol. 7, 2009, pp. 690-705.
31. P. Deepalakshmi and S. Radhakrishnan, “An ant colony-based multi objective quality
of service routing for mobile ad hoc networks,” EURASIP Journal on Wireless Com-
munications and Networking, Vol. 2011, 2011, pp. 1-12.
32. S. Samadi and H. Beigy, “An adaptive multipath ant routing algorithm for mobile ad
hoc networks,” International Journal of Computer and Electrical Engineering, Vol. 4,
2012, pp. 175-180.
33. G. Singh, N. Kumar, and A. K. Verma, “Ant colony algorithms in MANETs: A re-
view,” Journal of Network and Computer Applications, Vol. 35, 2012, pp. 1964-1972.
34. H. S. Urade and R. Patel, “Dynamic particle swarm optimization to solve multi-ob-
jective optimization problem,” Procedia Technology, Vol. 6, 2012, pp. 283-290.
35. A. Al-Ani and J. Seitz, “An approach for qos-aware routing in mobile ad hoc net-
works,” in Proceedings of IEEE International Symposium on Wireless Communica-
tion Systems, 2015, pp. 626-630.
36. V. V. Mandhare, V. R. Thool, and R. R. Manthalkar, “QoS routing enhancement using
metaheuristic approach in mobile ad-hoc network,” Computer Networks, Vol. 110,
2016, pp. 180-191.
37. D. Sarkar, S. Choudhury, and A. Majumder, “Enhanced-ant-AODV for optimal route
selection in mobile ad-hoc network,” Journal of King Saud University-Computer and
Information Sciences, Vol. 48, 2018, pp. 1-16.
38. M. K. Patel, M. R. Kabat, and C. R. Tripathy, “A hybrid ACO/PSO based algorithm
for QoS multicast routing problem,” Ain Shams Engineering Journal, Vol. 5, 2014,
pp. 113-120.

Priyanka Kumari was born on 28th January 1989. She re-

ceived the B.E degree in Information Technology from Radharaman
Engineering College (REC), Bhopal MP, India in 2011 and MTech
degree in Software System from Samrat Ashok Technological
Institute (S.A.T.I), Vidisha, M.P, India in 2014. She is currently a
Research Scholar in the Department of Computer Science and Engi-
neering, B.I.T., Mesra, Ranchi, India. Her current interests include
soft computing, computational intelligence, artificial intelligence
and computer network.
56 PRIYANKA KUMARI AND SUDIP KUMAR SAHANA

Sudip Kumar Sahana was born in Purulia West Bengal, India

on 8th October 1976. He received the B.E degree in Computer Tech-
nology from Nagpur University, India in 2001, and the M.Tech. de-
gree in Computer Science in 2006 from the B.I.T (Mesra), Ranchi,
India where he has done his Ph.D. (Engineering) in 2013. He is cur-
rently working as Assistant Professor in the Department of Com-
puter Science and Engineering, B.I.T., Mesra, Ranchi, India. His re-
search and teaching interests include soft computing, computational
intelligence, distributed computing and artificial intelligence. He has
authored numerous articles, research papers and books in the field of Computer Science
and assigned as editorial team member and reviewer for several reputed journals. He is a
lifetime member of Indian Society for Technical Education (ISTE), India and fellow of
IETE, India.