Ant Colony Optimization Based Modified Termite Algorithm (MTA) with Efficient Stagnation Avoidance Strategy for MANETs

International journal on applications of graph theory in wireless ad hoc networks and sensor networks
(GRAPH-HOC) Vol.4, No.2/3, September 2012
DOI : 10.5121/jgraphoc.2012.4204 39
ANT COLONY OPTIMIZATION BASED
MODIFIED TERMITE ALGORITHM (MTA)
WITH EFFICIENT STAGNATION
AVOIDANCE STRATEGY FOR MANETS
Sharvani G S, Dr. A G Ananth and Dr T M Rangaswamy
R V College Of Engineering
Mysore Road, Bangalore
sharvanigs@rvce.edu.in, antisro@yahoo.com, rangaswamytm@rvce.edu.in
ABSTRACT
Designing an effective load balancing algorithm is difficult due to Dynamic topology of MANET. To
address the problem, a load balancing routing algorithm namely Modified Termite Algorithm (MTA) has
been developed based on ant’s food foraging behavior. Stability of the link is determined based on node
stability factor ‘∆’. The stability factor “∆ “of the node is the ratio defined between the “hello sent” and
“hello replied” by a node to its neighbors. This also indicates the link stability in relation to other paths
towards the destination. A higher ratio of “∆” indicates that the neighbor node is more stable. Using this
concept pheromone evaporation for the stable node is fine tuned such that if the ratio “∆” is more, the
evaporation is slow and if “∆” is less the evaporation is faster. This leads to decreasing of the pheromone
content in an optimal path which may result in congestion. These paths can be avoided using efficient
evaporation technique. The MTA developed by adopting efficient pheromone evaporation technique will
address the load balancing problems and expected to enhance the performance of the network in terms of
throughput, and reduces End-to-end delay and Routing overheads.
KEYWORDS
Ant colony optimization, MANET, Node stability, Termite Algorithm, stagnation
1. INTRODUCTION
Data transfer on the move is highly essential and is the basic need in MANETs. Mobility of nodes
in MANETs affects the performance of routing protocols. Using greedy mechanism, the network
tends to choose a single path every time which leads to congestion. Designing an effective load
balancing algorithm is difficult due to dynamic topology of MANET. To address this problem, a
load balancing routing algorithm Modified Termite Algorithm (MTA) has been developed based
on ant’s food foraging behavior.
The MTA developed by adopting efficient pheromone evaporation technique will address load
balancing problems and will enhance the performance of the network in terms of throughput, and
reduces End-to-end delay and routing overheads.

40
1.1 Mobile Ad-hoc Networks (MANETs)
Wireless communications can be categorized into Infrastructure dependent and infrastructure
independent networks. The Cellular networks are infrastructure dependent where as Ad hoc
Wireless Networks (AWN) are capable of operating without any fixed infrastructure support.
MANET is one such category of AWN, which has a set of mobile agents, communicating with
each other with-in radio frequency range. Since radio frequency range is limited, the
communication traffic has to be relayed over several intermediate nodes. Therefore, MANET is
also known as Multi Hop ad hoc networks. The functionality of the nodes is not only to fulfill as a
host, but also as a router to forward packets to other nodes. Routing in MANET is a challenging
task due to characteristics like node mobility - which leads to dynamic Topology, Error prone
shared channel, channel contention and limited device availability. Hence, an efficient routing
protocol has to be developed which is capable of reducing control overhead and utilizes the
bandwidth efficiently [1] [2]. Ant Colony Optimization (ACO) is Swarm Intelligence (SI) based
routing algorithm proposed for high quality routing in MANET.
1.2 Ant Colony Optimization (ACO)
ACO was inspired by an algorithm called “Ant System (AS)” [3]. ACO deals with artificial
systems that are inspired from food foraging behavior of real ants, which can find optimal
solutions in highly dynamic environment. A major problem with ACO algorithm is “stagnation”.
This occurs when all ants try to follow the same path to reach the destination (since there is more
pheromone). This when applied in MANETs comes to a convergence state (equilibrium) and
attracts all the data packets to follow the same path, which leads to congestion.
To mitigate the above problem of ACO algorithms, different methodologies where adopted and
are categorized as follows.
• Privileged pheromone Laying
• Pheromone-heuristic control
• Pheromone Control.
The proposed work uses Pheromone control technique in the form of controlled evaporation for
efficient routing.
Wireless networks such as MANETS have many advantages compared to wired networks. In
MANETs the communication is not limited to a certain region (as in cellular region). MANETs
have some peculiar routing challenges. Swarm Intelligence (SI) based routing algorithms provides
an interesting analogy to the world of biology. SI based ACO algorithms provide interesting
solutions to network routing problems. ACO based routing in MANETs will enhance the
reliability and efficient packet delivery. They help in reducing control overhead due to their
inherent scalable feature. The similarity between ant and nodes, colony and Wireless network
helps to use ACO based routing in MANETs. Termite is one such ACO based algorithm which
performs better than the AdHoc On demand Distance vector (AODV).
2. Stagnation avoidance algorithms
Elitist Ant System (EAS) algorithm is a variation of ACO [4]. Here pheromone deposits are only
for the best path found. The major problem with this algorithm is early stagnation. Max-Min Ant
systems (MMAS) [5] overcome the early stagnation problem by adding an extra constraint which
says that the pheromone is bounded between maximum and minimum pheromone concentration.

41
Stagnation avoidance algorithm was developed while ACO found its application in Scheduling of
flexible manufacturing systems [6]. The methodology developed deals with quick convergence
and stagnation avoidance in which parameters of ACO such as evaporation co-efficient which
control the trail and its visibility are fine tuned. MMAS was further improved by adopting
Minimum Pheromone Threshold Strategy (MPTS)[7]. Here the bound between maximum and
minimum threshold is fine tuned based on performance of the network. Improved Lower Limits
for Pheromone Trails in ACO adopts improved estimates of the lower pheromone value. This
helps algorithms like MMAS (sets implicit pheromone trail limit) to avoid stagnation [8]. Another
work on stagnation avoidance for scheduling of real-time tasks [9] uses a non-preemptive
scheduling approach based on distance function as an extra parameter in the transition rule with
the pheromone information.
A framework for stagnation avoidance was developed while adapting ACO in TSP [10]. This
frame work evaluates each solution against the last “N” solutions globally constructed by ants. As
soon as “N” solutions are available, their moving average “A” is computed. New value of A is
compared with old value. If New A is lower than old, the trail level of the last solution’s moves is
increased, otherwise it is decreased.
An ACO algorithm with improved pheromone correction strategy for the Minimum Weight
Cover Problem (MWCP) [11] [12] explains that new Hybridization techniques. Corrections to the
pheromone trails are performed based on suspicion whether it is good or bad. It decrements the
pheromone trails very fast if it finds more suspicious elements. In Hybrid architecture, the age of
the ant decides whether the trail should be increased or not [13]. Another solution for stagnation
was found by using information entropy [14] which helped the algorithm to fine tune its
pheromone trails for local search.
3. Termite Hill Building process
Termite’s hill building process involves, collecting the pebbles spread over an area and
place it in one place. The cooperation between Termites will lead to constructing of a
huge hill (as big as 7 Feet). This is achieved by Termites moving on locally observed
pheromone trails (laid by other Termites).If no trails are found, it moves randomly in any
direction.
As shown in Fig 1 each Termite may carry only one pebble at a time and move observing
the trail laid by other Termites. If a pebble is found on the way it just puts down pebble
carried by it and both pebbles are infused with pheromone which then evaporates.
Termite’s inclination towards pheromone trail (positive feedback) leads to large piles of
pebbles forming a hill. At the same time evaporation (negative feedback) declines the
movement of Termite towards the direction in which hill building is not appropriate or
avoids small piles. It is essential that many individual work together, otherwise before the
pebbles added to the pile the evaporation of trail leads random walk by Termites without
forming any significant piles.
Analogous to the above example, each node in the Mobile wireless network is a Termite
Hill. As packets travel from source to the destination, it leaves pheromone for its source
and at the same time follows the pheromone for its destination through the MANET. The

42
pheromone is laid between nodes’ links which are communicating with each other. As in
Termite example, packets are inclined towards strong pheromone trails. Laying source
pheromone trail on the path increases the likely hood of packets travelling on the same
path (reverse path).
Fig 1 Flowchart of Termite Hill building process
4. Modified Termite Algorithm (MTA)
Termite algorithm is a routing protocol for MANET. The packets in the network follow
the principles of SI and results in emergent routing behavior. Packets are not forwarded to
the same neighbor from whom it received (to avoid cycles or loops). It is similar to ARA
in terms of routing data but varies with respect to route discovery and failure handling. It
is inspired by hill building procedure
.
Termites are biased towards the “pheromone” laid by other termites towards the
destination hill. This is very similar to packets in MANETs biased towards the optimal
path to reach the destination. Information about the network will be carried by data
packets which help in making probabilistic routing decision. This decision helps to find
maximum utility path in an emergent manner.
As shown in the context diagram Figure 2 the network gives the source and destination
node to the Termite Routing algorithm for which the data is to be routed.
Carrying a
Pebble?
Initialization
Termite’s Random
Walk
Found
Pebble?
Drop Pebble
Pick up
Pebble
No
Yes
Yes
No

43
Figure: 2 Routing in MANETs using Termite Algorithm
Figure: 3 Phases of MANETs using Termite Algorithm
The algorithm transmits destination, neighbor and pheromone value to and from the
routing table i.e. the routing table is updated. The output given is the route from the
source to destination through which the data is routed. The context diagram of Figure.2
gets segregated into several processes as shown in Figure 3. These processes work
together to achieve the desired result.
The first process here is the Route discovery which takes the source and destination node
from the network. It searches for a path from source to destination. Once the path is

44
discovered, the best path is maintained by sending data on that path by the Route
maintenance module. The third module is for Route Failure to detect for any link failure
and to find the new path.
This approach reduces control overhead and maximizes routing performance focusing on
the routing metric estimator “pheromone.” The pheromone is adjusted in such a way that
the best possible utility link is chosen for delivering the packets to the destination.
Fig 4‘S’ node neighborhood
4.1 Routing in MANETs with MTA
Before routing the network has to stabilize itself by sending hello packets (or through
sniffing) to its neighbors and build the routing or pheromone table. This helps them to
know their neighbors. Neighbor nodes must be carefully handled as they can forward/
originate packets. However, they always give preference for packets to be forwarded.
Thus, the neighbor row must be declared decayed before the neighbor node can be
removed from the pheromone table. If a neighbor is determined to be lost by means of
communications failure, the neighbor row is simply removed from the pheromone table.
Similarly a decayed column indicates that no traffic has been seen which was sourced by
that node.
4.1.1 Routing table in MTA
Termite’s pheromone table is analogous to routing table. Each node in the network
maintains a pheromone table which keeps track of percentage of pheromone in each
neighbor link. The table consists of row representing the neighbors and columns
correspond to destinations (except for node itself). Table size is dynamic and varies based
on the destinations heard recently. An entry of node ‘n’ in the table is referenced by P
n
d
i,
where N
n
i ∈ is the neighbor and‘d’ denotes the destination where D
n
d ∈ , in other
words, P
n
d
i,
is the percentage of pheromone from a node‘d’ on the link with neighbor ‘ ‘i
at node ’n’.
S
N1
N2
N3
D
Neighborhood

45
4.1.2 Pheromone update/Decay
Pheromone is updated only by data packets as compared to other ACO based algorithm in
which all forward ants, backward ants and data packets update pheromone concentration.
When a packets arrives at a node ‘n’ from source s and a previous hop ‘p’, the new
pheromone concentration entry for P
n
s
p ,
for the source of the packet is increased by
constant pheromone value ‘ γ ’ equation as shown in Eqn (1) the.
P
n
s
p ,
← P
n
s
p ,
+ γ ………………… Eqn (1)
Pheromone decay is an important factor for efficient packet delivery, if not handled
properly, it leads to stale entries in the network which lead to packet drops or it might
lead to improper load in the network due to data packets travelling on the optimal path
and increasing the pheromone concentration. This tells other packets to travel blindly not
knowing whether the link is overloaded or not.
Pheromone values in the routing table will be decayed periodically as shown in Eqn (2)
by multiplying it with decay factor, e-τ
where τ ≥0 and is a global parameter. There should
be a balancing act between the pheromone decay process i.e. neither too quick nor slow.
The nominal pheromone decay interval is one second which is called as decay period.
N
n
i ∈
∀ , D
n
d ∈
∀ , P
n
d
i,
← P
n
d
i,
.e-τ
......................... Eqn (2)
If all the pheromone for a particular node (as it has not received any packets for a long
time) decays, the row for that is removed from the routing table. However, column entry
cannot be deleted directly if it is also neighbor node. Pheromone bounds like max, min
and initial pheromone values (prevents extreme difference in pheromone value) in the
routing table plays a major role in stagnation problems. Table updates new column and
row for the neighbor as it can also be a destination. When a packet is received from
unknown source then the new entry for that node is created on the routing table. If the
nodes are just destinations then only the column are created.
For example as shown in Fig 4, Source ‘S’ considers N1, N2 and N3 as neighboring
nodes since they are within the transmission range. Once the neighboring nodes are
known then the routing table is updated as shown in Table 1(route discovery also helps in
updating the routing table).When node ‘S wants to communicate with destination ‘D’ it
checks for the highest trail for destination ‘D’ and then forwards the packet.
TABLE 1 ROUTING TABLE AT NODE ‘S’
D
N1 0.18
N2 0.27
N3 0.01
DESTINATION NODE
N
E
I
G
H
B
O
R

46
4.2 Efficient Stagnation avoidance Technique
Out of the many decay functions analyzed, the exponential decay techniques in ACO are
more applicable for dynamic networks like MANETs. The limitation of this technique is
pheromone evaporates (uniformly) very quickly after some time leaving no traces for the
data transmission. Hence, fresh route discovery is established to know the destination
address. This causes more control overhead and makes bandwidth inefficient. To deal
with this problem, controlled exponential evaporation is adopted which fine tunes
evaporation based on the stability factor ‘∆’ of the node .Each node in the network also
maintains the Node stability Table for all its neighboring nodes as shown in TABEL2.
This table helps in finding more stable nodes.
TABLE 2 NODE STABILITY FACTOR OF NEIGHBOR NODES AT NODE ‘S’
Evaporation is fine tuned based on the stability factor of the node in the vicinity of the
node that forwards the packet. During packet forwarding time, each node calculates the
‘hello’ sent and ‘hello’ received by its neighboring nodes and calculates the stability ratio
shown in Eqn (3)
Ratio = (Hello Replied / Hello Sent) * 100 ……………………….Eqn (3)
With the help of ratio calculated, evaporation is controlled for each neighboring node by
fine-tuning the decay speed like decaying fast for unstable nodes and decaying little
slower for stable nodes .
Exponential Decay may result in faster lost of pheromone trail traces. Pheromone trail is
controlled by adding an extra parameter ‘∆’ to the exponential decay factor.
0.003 If Ratio is > 75 %
Where ∆ = 0.002 If Ratio is ≤75 % and <55%
0.001 If Ratio is ≤ 55 %
This technique helps the MANETs to overcome stagnation problem, it also reduces the
control overhead and have efficient packet delivery ratio. ‘∆’ is tunable based on the
application and the node mobility as shown in Eqn 4. This tells with what percentage the
link should decay its pheromone (tunable). Table 3 describes pheromone decay factor vs
ratio.
Ratio= Hello sent / Hello replied
( %)
N1 55
N2 78
N3 24
STABILITY FACTOR
N
E
I
G
H
B
O
R

47
P
n
d
i,
← P
n
d
i,
.e-(τ+∆)
.......................Eqn (4)
If Ratio
≥ 75 %
If Ratio
< 75 % and ≥ 55%
If Ratio
< 55 %
Decay Factor e-(τ+∆)
( )
003
.
0
-
e +
τ
( )
002
.
0
-
e +
τ
( )
001
.
0
-
e +
τ
TABLE 3 DECAY FACTOR FOR DIFFERENT RATIO
Exponential Decay may result in faster lost of pheromone trail traces. Pheromone trail is
controlled by adding an extra parameter ‘∆’ to the exponential decay factor. Decay
should be efficiently handled (at an interval of one second); otherwise a high decay rate
will quickly reduce the amount of remaining pheromone, while a low value will degrade
the pheromone slowly.
A source ready to forward packet to the destination, takes packet forwarding equation’s
help to find out the next hop neighbor. This formula (as shown in Eqn (5)) updates
pheromone value for that destination‘d’ on each outgoing link ‘i’. The random selection
of next neighbor is performed with the help of packet forwarding equation. Extra care
must be taken by not forwarding packet to the previous hop ‘p’.
( )
( )
∑ −
∈
+
+
=
p
n
S
j
F
F
K
P
K
P
P
n
d
j,
n
d
i,
n
d
i, ………………………….Eqn (5)
Where F (pheromone sensitivity) and K (pheromone threshold) are tunable based on their
sensitivity to amount of pheromone values.
3. RESULTS AND DISCUSSION:
The exponential pheromone decay technique based on node stability is adopted which makes the
MANET efficient in terms of control overhead, bandwidth, packet delivery etc When applied on
Termite algorithm, proved to be better in terms of throughput, end-to-end delay and routing
overhead.
Modified Termite Algorithm (MTA) based on ACO has been implemented on MANET. The
primary metrics considered for the algorithm are Throughput, End-to-End delay and routing over
head against the packet size. A series of tests are conducted to check the feasibility of MTA as an
effective mechanism for MANETs based on the community based mobility model. The behavior
of success rates with the number of termites (nodes) has been studied. It is observed that only
after 30 nodes, the algorithm shows a sharp increase in the success rate (considering 5 m/s
mobility speed and pause time of 2 secs). It is considered as appropriate to choose the number of
termites (Nodes) ~ 30 for determining the throughput, end-to-end delay and control overheads for
the MANET.

48
The ACO based MTA has been implemented on the MANET to determine the variation of
throughput with the increase of packet sizes between 1-6 Kb. Figure 1 shows the Throughput
variation for different packet sizes for MTA . In the same figure the throughput variations for the
termite algorithm which does not incorporate efficinet pheromone decay technique is also shown
for comparison.
Fig 1 Throughput (Mbps/sec) Vs Packet size (Kb) Analysis
Figure 1 clearly demonstrates that the MTA based MANET shows continuous increase in
throughput ~4.0 times with the packet size, where as the throughput for termite algorithm remains
low < 1.0 and does not show any significant variations with the packet size. This suggests that
the MTA algorithm implemented on a MANET with 30 nodes indicates large improvement in the
throughput ~75% with the increase of packet size as compared to termite algorithms
The MTA implemented on MANET has been used to determine the end–to-end delay with the
increase of packet sizes in the range of 1-6 Kb. Figure 2 shows the end to end delay variations for
different packet sizes. In the same figure the end to end delay variations for the termite algorithm
which does not incorporate efficinet pheromone decay technique is also shown for comparison.
Fig 2 End to End Delay (msec) Vs Packet size (Kb) Analysis

49
It is seen form the figure 2 that the end to end delay for termite algorithm is ~ 1.53 times increase
in time delay with the packet size. Where as the end to end delay variation for MTA algorithm <
0.3 and remains almost constant and do not show any significant variations with the packet size
This suggests that the MTA algorithm implemented on a MANET with 30 nodes shows reduction
in the end–to-end delay ~23% with the packet size as compared to termite algorithms.
The MTA implemented on MANET has been used to determine the routing over head changes
with the increase of packet sizes between 1-6 Kb. Routing overhead is the ratio of number of
control packets required for delivery of data .Figure 3 shows the routing over head changes
obtained using MTA algorithm for different packet sizes. In the same figure the routing over
head changes for the termite algorithm is shown for comparison.
Fig 3 Routing Overhead (packets) Vs Packet size (Kb) Analysis
The figure 3 clearly demonstrates that for termite algorithms the Routing overheads ~ 1.15 times
increase with the packet size. Where as the routing overheads increases for MTA algorithm ~1.5
times with the packet size This suggests that the MTA algorithm implemented on a MANET
with 30 nodes shows reduction in the overheads ~13% with the packet size as compared to
termite algorithms suggesting that MTA algorithms shows improvement in routing overheads
compared to termite algorithms.
4. CONCLUSION
From the analysis carried out, the following conclusions are drawn
1. The stagnation problems can be overcome by fine tuning of the Pheromone concentration
based on node stability factor.
2. The MTA implemented on MANET is found to perform better for 40 nodes between as
the success rates are found to be maximum in this range.
3. The MTA implemented on MANET with fine tuning of pheromone concentration shows
significant increase in the throughput ~75% with the increase of packet size from 1-6 Kb
compared to termite algorithms

50
4. The MTA implemented on MANET shows significant decrease in End-to-end time delay
23% with the increase of packet size from 1-6 Kb compared to termite algorithms
5. The MTA implemented on MANET shows a small decrease ~13% in the routing
overheads with the increase of packet size from 1-6 Kb and indicates improvement in
routing overheads compared to termite algorithms
REFERENCES
[1]. C S R.Murthy and B S Manoj, “ Ad Hoc Wireless Networks : Architecture and protocols” prentice
Hall PTR, 2004
[2]. Elang Huang, Wenjun Hu Jon, Crowcroft Lan, “ Towards commercial Mobile Ad Hoc Network
applications: A Radio Dispacth System”, MobiHOC May, 2005,Illinois, USA.
[3]. March Dorigo,” Ant Colony Optimisation and swarm intelligence”, 4th
International workshop,
ANTS, 2004, Brussels, Belgium, Sep 2004( Proceedings in LNCS, Berlin:Springer)
[4]. Ducatelle F, Di Caro G and Gambardella L M, “Principles and applications of Swarm Intelligence for
adaptive routing in telecommunications networks”, Swarm Intelligence, 2010.
[5]. T Stutzle, H H Hoos,”Max-Min Ant system” Future Generation Computing Syst , PP 889-914,2000.
[6]. R Kumar, M K Tiwari and R Shankar, “Scheduling of flexible manufacturing systems: an ant colony
optimization approach”, Proceedings Instn Mech Engrs Vol 217, Part B: J Engineering Manufacture,
,pp 1443-1453,2003.
[7]. Kuan Yew Wong, Phen Chiak See, “ A New minimum pheromone threshold strategy(MPTS) for
Max-min ant system “, Applied Soft computing, Vol 9, pp 882-888 , 2009
[8]. David C Mathew, “ Improved Lower Limits for Pheromone Trails in ACO”, G Rudolf et al(Eds),
LNCS 5199, pp 508-517, Springer Verlag, 2008.
[9]. Laalaoui Y, Drias H, Bouridah A and Ahmed R B, “ Ant Colony system with stagnation avoidance
for the scheduling of real time tasks”, Computational Intelligence in scheduling, IEEE symposium,
pp 1-6,2009.
[10]. E Priya Darshini, “ Implementation of ACO algorithm for EDGE detection and Sorting Salesman
problem”,International Journal of Engineering science and Technology, Vol 2, pp 2304-2315, 2010
[11]. Alaa Alijanaby, KU Ruhana Kumahamud, Norita Md Norwawi, “Interacted Multiple Ant Colonies
optimization Frame work: an experimental study of the evaluation and the exploration techniques to
control the search stagnation”, International Journal of Advancements in computing Technology Vol
2, No 1, pp 78-85,March 2010 .
[12]. Raka Jovanovic and Milan Tuba, “An ant colony optimization algorithm with improved pheromone
correction strategy for the minimum weight vertex cover problem”, Elsvier, Applied Soft Computing,
PP 5360-5366, 2011.
[13]. Priyanka Sharma, Dr K Kotecha, “ Optimization in stagnation avoidance of ACO based routing of
Multimedia Traffic over Hybrid MANETs”, International Journal of computer science and
technology, IJCST, Issue 2, ISSN: 2229-4333(print), 0976-8491(online), 2011
[14]. Zar Ch Su Hlaing, May Aye Lhine, “An Ant Colony Optimization Algorithm for solving Traveling
Salesman Problem”, International Conference on Information Communication and management
(IPCSIT), Vol, 6, pp 54-59, 2011.
[15]. Martin Roth and Stephen Wicker “Termite: Emergent Ad-Hoc Networking " Wireless Intelligent
Systems Laboratory School of Electrical and Computer Engineering Cornell University Ithaca, New
York 14850 USA
[16]. Sharvani G S, Dr. Ananth G and Dr T M Rangaswamy, "Analysis of different pheromone decay
techniques for ACO Based routing in Ad Hoc wireless Networks “( communicated).

Ant Colony Optimization Based Modified Termite Algorithm (MTA) with Efficient Stagnation Avoidance Strategy for MANETs

More Related Content

Similar to Ant Colony Optimization Based Modified Termite Algorithm (MTA) with Efficient Stagnation Avoidance Strategy for MANETs (20)

More from GiselleginaGloria (20)

Recently uploaded (20)

Ant Colony Optimization Based Modified Termite Algorithm (MTA) with Efficient Stagnation Avoidance Strategy for MANETs