International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No.

2, April 2011

ENHANCING THE PERFORMANCE OF THE DNDP

ALGORITHM

Hussein Al-Bahadili1, Ghassan Issa1, Alia Sabri2

1
Faculty of Information Technology, Petra University, Jordan
2
Faculty of Information Technology, Applied Science University, Jordan
hbahadili@uop.edu.jo, gissa@uop.edu.jo, a_sabri@asu.edu.jo
ABSTRACT
A novel dynamic noise-dependent probabilistic (DNDP) route discovery algorithm was recently
developed to enhance the performance of the dynamic probabilistic algorithm in noisy mobile ad hoc
networks (MANETs). In this algorithm, the mathematical model for calculating node retransmission
probability (pt) is calculated as a function of two independent variables: number of first-hop neighbors
(k) and probability of reception (pc). The model also shows another independent variable, namely, the
maximum retransmission probability that can be assigned to the transmitting node (pt,pcmin), which is
assumed to be a fixed value. In this paper, we propose a new mathematical model for calculating pt. In
this new model pt,pcmin is calculated as a function of k. The performance of the DNDP algorithm using
fixed and k-dependent pt,pcmin is evaluated through simulations. The simulation results showed that the
new model enhances the performance of the DNDP algorithm as it significantly reduces the number of
retransmissions at an insignificant reduction in the network reachability.

KEYWORDS
MANETs, dynamic routing protocols, route discovery, flooding optimization algorithms, probabilistic
algorithm, the DNDP algorithm.

1. INTRODUCTION
A mobile ad hoc network (MANET) is defined as a collection of low-power wireless mobile
nodes forming a temporary network without the aid of any established infrastructure or
centralized administration [1, 2]. A data packet in MANET is forwarded to other mobile nodes
on the network through reliable and efficient routing protocols, which are usually implemented
in software as part of the network layer [3]. Dynamic routing protocols (DRPs), such as: ad hoc
on-demand distance vector routing (AODV) [4], dynamic source routing (DSR) [5], zone
routing protocol (ZRP) [6], and location-aided routing (LAR) [7], are widely used routing
protocols in MANETs.
The DRPs consist of two phases: (i) route discovery, which is initiated when a node (source)
desires to send a data packet to some node (destination) and does not have a valid route to that
destination, and (ii) route maintenance, which is periodically initiated to maintain existing route;
and if the route is broken for any reason, then the source either finds and uses other recognized
route on its routing table or if it cannot find a route, it initiates a new route discovery procedure
[3]. It is well approved that the cost of information exchange during route discovery is higher
than the cost of point-to-point data forwarding after the route is discovered. Thus, significant
efforts have been put forward by researchers to minimize the cost of route discovery [8].
In route discovery, the source initiates and broadcasts to its neighbours a small packet called
route request packet (RREQ), which carries information on the source and destination
addresses, packet sequence number (ID), and specific lifetime, i.e., time-to-live (TTL). Each
recipient node checks to see if it is the destination; if not, then it appends its address to the
payload of the RREQ and forwards it to its neighbours, and so on until the expiration of the
DOI : 10.5121/ijwmn.2011.3210 113
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

RREQ. If the RREQ reaches its destination, the destination sends a route reply packet (RREP)
back to the source through the route from which it first received that particular RREQ [3, 5]. If
the RREQ expires before reaching its destination, the node at which it expires, sends a route
error packet (RERR) back to the source, once again through the route from which it first
received that particular RREQ. As result of this route discovery failure, the source initiates a
new RREQ and continues to do so until a route is found.
Pure flooding is one of the simplest and reliable mechanisms proposed in the literature for route
discovery in MANETs [1, 9]. The main drawback of pure flooding is that it is very costly where
it costs n transmissions in a network of n reachable nodes. In addition, it results in serious
redundancy, contention, and collisions in the network; such a scenario has often been referred to
as the broadcast storm problem (BSP) [10].
To eliminate the effects of BSP, a variety of flooding optimization algorithms have been
developed, such as: probabilistic [1, 2, 11], multipoint relaying (MPR) [12, 13],
counter/distance/location/cluster -based [3, 7, 10, 14] algorithms. They all try to minimize
contention and collisions by limiting the number of retransmissions. As the number of
retransmissions is decreased, the bandwidth and power are saved and contentions and collisions
are reduced, and consequently the overall network performance is improved.
Probabilistic algorithm has been widely-used for route discovery in MANETs [1, 2]. However,
it has been demonstrated that the performance of the probabilistic algorithm is severely suffered
in presence of noise due to increases packet-loss (data or RREQ packets), and consequently
reduces the overall network performance [15, 16].
A dynamic noise-dependent probabilistic (DNDP) algorithm was proposed by Al-Bahadili and
Sabri in [17] for route discovery in noisy MANETs. In this algorithm, the retransmission
probability of the transmitting node (pt) is modeled as a function of two independent variable;
these are: the number of first-hop neighbors (k) and the probability of reception (pc). The model
also shows another independent variable, namely, the maximum retransmission probability that
can be assigned to the transmitting node (pt,pcmin), which is assumed to be a fixed value.
In this paper, we propose a new mathematical model for calculating pt. In this new model, pt,pcmin
is calculated as a function of k. The new model is implemented on the MANET simulator
(MANSim) [18], and the performance of the DNDP algorithm using fixed and k-dependent
pt,pcmin is evaluated through simulations. The simulation results showed that the new model
enhances the performance of the DNDP algorithm as it reduces the number of retransmissions at
an insignificant reduction in the network reachability.
This section provides an introduction to the general domain of this paper. The rest of the paper
is organized as follows. Related work is discussed in Section 2. The dynamic probabilistic and
the DNDP algorithms are discussed in Sections 3 and 4, respectively. The proposed model is
described in Section 5. Simulation results are presented and discussed in Section 6. Finally, in
Section 7, conclusions are drawn and recommendations for future work are pointed-out.

2. PREVIOUS WORK
This section reviews some of the most recent and related work on probabilistic flooding in both
noiseless and noisy MANETs. Probabilistic algorithm was first used by Haas et. al. [19] for
route discovery in ad hoc networks, and they called it a gossip-based route discovery
(GOSSIP1) approach. They used a predefined pt to decide whether or not a node forwards the
RREQ packets. GOSSIP1 has a slight problem with initial conditions. If the source has
relatively few neighbors, there is a chance that none of them will gossip, and the gossip will die.
To make sure this does not happen, Haas et. al. later developed a modified protocol, in which
they gossip with pt=1 for the first h hops before continuing to gossip with pt<1. Their results
showed that they can save up to 35% message overhead compared to simple flooding.
114
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

S. Tseng et. al. [10] investigated the performance of the probabilistic flooding for various
network densities in noise-free environment. They presented results for three network
parameters, namely, reachability, saved rebroadcast, and average latency, as a function of pt and
network density. Sasson et. al. [20] suggested exploring algorithms in which nodes would
dynamically adjust their pt based on local topology information. Because in their work they
made the assumption that all nodes possess the same transmission range, they suggested another
potential area for study which is to modify pt according to the node radio transmission range.
Kim et. al. [21] introduced a broadcasting scheme in which a node dynamically adjusts its pt
according to its additional coverage area. The additional coverage is estimated by the distance
from the sender. The simulation results showed this scheme generates fewer rebroadcasts than
pure flooding approach. It also incurs lower broadcast collision without sacrificing high
reachability.
Scott and Yasinsac [2] presented a dynamic probabilistic solution that is appropriate to solving
BSPs in dense mobile networks. The approach can prevent broadcast storms during flooding in
dense networks and can enhance comprehensive delivery in sparse networks.
Barret et. al. [22] introduced a probabilistic routing protocol for sensor networks. In this
protocol, a sensor decides to forward a message with pt that depends on various parameters,
such as the distance of the sensor to the destination, the distance of the source sensor to the
destination, or the number of hops a packet has already traveled. They proposed two protocol
variants and compared the new methods to other probabilistic and deterministic protocols. The
results showed that the multi-path protocols are less sensitive to misinformation, and suggest
that in the presence of noisy data, a limited flooding strategy will actually perform better and
use fewer resources than an attempted single-path routing strategy, also parametric probabilistic
protocols outperforms other protocols. The results also suggested that protocols using network
information perform better than protocols that do not, even in the presence of strong noise.
Viswanath and Obraczka [23] developed an analytical model to study the performance of plain
and probabilistic flooding in terms of its reliability and reachability in delivering packets. Their
results indicated that probabilistic flooding can provide similar reliability and reachability
guarantees as plain flooding at a lower overhead.
Zhang and Agrawal [24] proposed a probabilistic scheme that dynamically adjusts pt as per
node distribution and node movement. The scheme combines between probabilistic and counter-
based approaches. They evaluated the performance of their scheme by comparing it with simple
flooding and fixed probabilistic algorithms. Simulation results showed that the new scheme
performs better than the two algorithms.
Abdulai et. al. [25] studies the performance of the AODV protocol over a range of possible pt.
They focused on the route discovery part of the routing algorithm, they modified the AODV
implementation to incorporate pt; the RREQ packets are forwarded in accordance with a
predetermined pt. Simulation results showed that setting efficient pt has a significant effect on
the general performance of the protocol. The results also revealed that the optimal pt for
efficient performance is affected by the prevailing network conditions such as traffic load, node
density, and node mobility. During their study they observed that the optimal pt is around 0.5 in
the presence of dense network conditions and around 0.6 for sparse network conditions.
Bani-Yassein et. al. [1] proposed a dynamic probabilistic flooding algorithm in MANETs to
improve network reachability and saved rebroadcast. The algorithm determines pt by
considering the network density and node movement. This is done based on locally available
information and without requiring any assistance of distance measurements or exact location
determination devices. The algorithm controls the frequency of rebroadcasts and thus might
save network resources without affecting delivery ratios.

115
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

Abdulai et. al. [26] proposed two probabilistic methods for on-demand route discovery, that is
simple to implement and can significantly reduce the overhead involved in the dissemination of
RREQs. The two probabilistic methods are: the adjusted probabilistic (AP) and the enhanced
adjusted probabilistic (EAP) which address the broadcast storm problem in the existing OADV
routing protocols.
Bani Yassein et. al. [9] combined probabilistic and knowledge based approaches on the AODV
protocol to enhance the performance of existing protocol by reducing the communication
overhead incurred during the route discovery process. The simulation results revealed that
equipping AODV with fixed and adjusted probabilistic flooding helps to reduce the overhead of
the route discovery process whilst maintaining comparable performance levels in terms of saved
rebroadcasts and reachability as achieved by conventional AODV. Moreover, the results
indicated that the adjusted technique results in better performance compared to the fixed one.
Khan et. al. [27] proposed a coverage-based dynamically adjusted probabilistic forwarding
scheme and compared its performance with simple and fixed probabilistic schemes. The
proposed scheme keeps up the reachability of simple flooding while maintaining the simplicity
of probability based schemes.
Hanash et. al. [28] proposed a dynamic probabilistic broadcast approach that can efficiently
reduce broadcast redundancy in MANETs. The algorithm dynamically calculates pt according to
k. They compared their approach against simple flooding approach, fixed probabilistic
approach, and adjusted probabilistic flooding by implementing them in a modified version of
the AODV protocol using GloMoSim. The simulation results showed that broadcast redundancy
can be significantly reduced through their approach while keeping the reachability high.
Al-Bahadili [29] developed a new retransmission probability adjusting model, in which the
neighborhood densities are divided into three regions (low, medium, and high). The
performance of the new model was evaluated and compared with pure and other probabilistic
algorithms. The model enhances the performance of probabilistic broadcast by reducing the
number of transmissions while keeping almost the same network reachability.
Al-Bahadili and Kaabned [15] investigated the effect of noise-level on the performance of the
probabilistic algorithm in MANETs. They investigated the effect of node density, node average
speed, radio transmission range, pt, and pc on number of retransmissions, duplicate reception,
average hop count, and reachability. Their results showed that the performance of the network is
severely suffered as pc increases, i.e. the noise-level increases.

3. THE DYNAMIC PROBABILISTIC ALGORITHM

In probabilistic broadcast, a node retransmits the packet with a certain pt and with probability
(1−pt) it discards the packet [1, 2]. A node is allowed to retransmit a given RREQ only once,
i.e., if a node receives a RREQ, it checks to see if it has retransmitted it before, if so then it just
discards it, otherwise it performs its probabilistic retransmission check. Nodes usually can
identify the RREQ through its sequence number. The source node pt is always set to 1, to enable
it initializing a RREQ. On other hand, the pt of the intermediate nodes (all nodes except the
source and destination) can be set by using one of the following approaches:
(1) Static approach in which a pre-determined pt is set for each node on the network and it
can be expressed as: pt=Pt, where Pt is a constant value (0<Pt≤1). Pt =1 for pure
flooding.
(2) Dynamic approach in which each node on the network locally calculates its pt as a
function of k (pt(k)), where pt(k) could be any linear/nonlinear function of k.

116
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

In this paper, we mainly concern with dynamic probability, therefore, in the next section, we
shall provide a description of an efficient and flexible function that will be used for dynamically
calculate pt(k) for intermediate nodes on the network.
3.1 Calculation of pt(k)
Many functions have been developed for calculating pt(k) [1-2, 19-21, 24, 29]. However, in this
paper, the discrete function presented in [29] is used for calculating pt(k). This function is shown
in Figure 1. It demonstrates a satisfactory performance when used in various network conditions
and it is mathematically expressed as:

(1)

Where pmin and pmax are the minimum and maximum retransmission probabilities, N1 is the
number of nodes at or below which pt(k)=pmax, N2 is the number of nodes at or above which
pt(k)=pmin, and p1 and p2 are the retransmission probabilities assign to intermediate nodes when
they have N1+1 and N2-1 first-hop neighboring nodes. p1 and p2 should lie between pmax and pmin,
and also p1 is always ≥p2.

pmax
Retransmission probability (pt)

Linear distribution
p1 pt=Linear function
Constant distribution
pt pt=pmed

p2 pmin

N1 N2 n
k
Number of first-hop neighbors (k)
Figure 1. The k-dependent retransmission probability (pt(k))
In general, selection of a satisfactory distribution within the interval [N1+1, N2-1] and the values
of pmax, pmin, p1, p2, N1, and N2 depend on a number of factors and need to be carefully selected.
In this work, the values in [29] are used, which are 0.8, 0.5, 0.8, 0.5, 4, and 15 for pmax, pmin, p1,
p2, N1, and N2.
Figure 2 outlines the main phases of the dynamic probabilistic algorithm in a noisy
environment. To accommodate the effect of noise, it can be seen in Figure 2 that when the node
is within the ratio transmission range of the transmitting node, a test is performed to find-out
whether the RREQ packet is successfully delivered or not. This is tested by generating a random
number ξ1 and comparing it with pc. If ξ1≤pc, then the packets is successfully delivered to the
receiving node, otherwise, it is undelivered [15, 16].
4. THE DNDP ALGORITHM
The DNDP algorithm was proposed to enhance the performance of dynamic probabilistic
algorithm in noisy MANETs. In this algorithm, instead of calculating pt as a function of k only,
pt is determined locally by the retransmitting nodes considering both k and pc (pt(k, pc)) as:
pt(k, pc) = pt(k) + pt(pc) (2)

117
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

Where pt(pc) is the noise-dependent pt.

Dynamic Probabilistic Algorithm
If (the receiving node is within the radio transmission range of the sender) Then
ξ 1=rnd() {ξ 1 some random number between 0 and 1}
If (ξ 1<= pc) Then {Reception random test}
Update the node reception index
If (The node has not retransmitted the received packet before) Then
ξ 2=rnd() {ξ 2 some random number between 0 and 1}
pt=function_pt()
If (ξ 2≤pt) Then
Retransmit the received packet
Update the node retransmission index
End if
End if
End if
End if
Function_pt() {Determining pt}
If (Static probability) Then
pt=constant value
Else (Dynamic probability)
pt=f(k, pc)
End If
Figure 2. The dynamic probabilistic algorithm

4.1 Calculation of pt(k, pc)

It can be seen from Eqn. 2 that in order to calculate pt(k, pc), we need to calculate pt(k) and
pt(pc). pt(k) can be calculated using Eqn. 1, and in this section we discuss the calculation of
pt(pc). The following main constraints should be considered when calculating pt(pc),:
(1) The value of pt(pc) should be ≥0 and ≤1-pt(k), so that pt(k, pc) will always be ≤1.
(2) The value of pt(k, pc) lies between pt(k) and pt,pcmin, which is a pre-adjusted maximum
allowable pt at a certain minimum value of pc (pc,min).
(3) The value of pt,pcmin should be ≤1 and ≥pt(k).
Considering the above constraints, pt(pc) can be calculated as:
pt(pc) = α (pt,pcmin - pt(k)) (3)
Where α is called the noise-correction factor, which is a function of pc, and has a value that lies
between 0 and 1. Substituting Eqn. 3 into Eqn. 2 yields:
pt(k, pc) = pt(k) + α (pt,pcmin - pt(k)) (4)
In [17], the value of α is calculated using the following equation:
1 − pc
α= (5)
1 − pc ,min

Now, substituting Eqn. 5 into Eqn. 4 yields the following general equation for pt(k, pc):
1 − pc
pt ( k , pc ) = pt ( k ) + ( pt , pc min − pt (k )) (6)
1 − pc ,min

It can be seen from Eqn. 6 that pt(k, pc) depends on pt(k), k, pc, pc,min, and pt,pcmin.
In a noiseless environment pc=1, α=0, and consequently pt(k, pc)=pt(k), i.e., pt is a function of k
only. In a noisy environment, when pc=pc,min, then α=1 and pt(k, pc)=pt,pcmin. If pc is any value
118
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

between pc,min and 1, then pt(k, pc) varies between pt(k) and pt,pcmin depending on pc. According to
the above discussion, pt(k, pc) always lies between pt(k) and pt,pcmin as shown in Figure 3.

Line along which pc=pc,min

pt,pcmin

Retransmission probability pmax

pc
pt(k, pc)

pt(k) Line along which

pmin

0
0 Nmin Nmax
k n-
Number of first-hop neighbors (k)
Figure 3. Variation of pt(k, pc) (solid area) with k (constant pt,pcmin)

5. ENHANCING THE DNDP ALGORITHM

In Eqn. 5, pt,pcmin is assumed to be a constant value, and as it can be seen in Figure 3 that the
computed pt is increased considerably with decreasing pc. This may result in a significant
increase in the number of retransmission, especially at high k. In order to alleviate the impact of
this increase in pt and to provide adjustable measure to control that increase, we introduce the
following solution. The solution is simply based on using a k-dependent distribution for
calculating pt,pcmin, i.e., pt,pcmin can be set as a function of k (pt,pcmin(k)) as shown in Figure 4.

pt,pcmin
Line along which pc=pc,min

pmax pt,pcmin(k)=f(k)
Retransmission probability

pc
pt(k, pc)

pt(k) pt,pcmin2

Line along which

pmin pc=pc,max=1

0
0 Nmin Nmax
k n-1
Number of first-hop neighbors (k)

Figure 4. Variation of pt(k, pc) (solid area) with k (pt,pcmin(k))

This distribution can be discrete similar to pt(k) or continuous distribution, such as linear,
parabolic, exponential, etc. Thus, Eqn. 6 can be re-written as:

119
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

(7)

Mathematical representation of the linear, parabolic, and exponential distributions of pt,pcmin(k)

can be given as follows:
pt , pc min 2 − pt , pc min1
Linear distribution: pt , pc min (k ) = pt , pc min1 + k (8)
n −1
pt , pc min 2 − pt , pc min1
Parabolic distribution: pt , pc min ( k ) = pt , pc min1 + k2 (9)
( n − 1) 2

pt , pc min 2 − pt , pc min1
Exponential distribution: pt , pc min (k ) = pt , pc min1 + (1 − e − β k ) (10)
1 − e − β ( n −1)
Where pt,pcmin1 and pt,pcmin2 are the maximum allowable pt in presence of noise at k=0 and k=n-1,
respectively, and β is a constant (β≤1). It can be deduced from Figure 4 that due to the reduced
size of the shaded area when compared with Figure 3, pt(k, pc) can be tuned in a controlled way
by choosing a proper distribution for pt,pcmin(k) for the sake of achieving optimum performance
in terms of network reachability and number of redundant retransmissions.
6. RESULTS AND DISCUSSION
In this section, we investigate the effect of pt,pcmin(k) on the performance of the DNDP algorithm
through a number of simulations using the MANET simulator (MANSim) [18]. Also, in this
section, the performance of the DNDP algorithm with different distribution for pt,pcmin(k) is
compared against the performance of pure flooding, fixed and dynamic probabilistic algorithms,
and the DNDP algorithm with fixed pt,pcmin.
A number of network performance measures are calculated using MANSim, such as: network
reachability (RCH), number of retransmission (RET), average duplicate reception (ADR),
average hop count (AHP), saved rebroadcast (SRB), and disconnectivity (DIS). These
parameters are recommended by the Internet Engineering Task Force (IETF) group to judge the
performance of the flooding optimization algorithms. Definition of these parameters can be
found in [13].
This paper presents results for two parameters only, these are: RCH and RET. RCH is defined
as the average number of reachable nodes by any node on the network normalized to n; or the
probability by which a RREQ packet successfully delivered from source to destination node.
RET is defined as the average number of retransmissions normalized to n. The input parameters
for these simulations are given in Table 1.

Table 1 - Input parameters.

Parameters Values
Geometrical model Random node distribution
Network area 600x600 m
Number of nodes (n) 100 nodes.
Transmission radius (R) 100 m
Average node speed (u) 5 m/sec
Probability of reception (pc) From 0.5 to 1.0 in step of 0.1
Simulation time (Tsim) 1200 sec
Pause time (τ) τ = 0.75*(R/u) = 15 sec
Size of the mobility loop (nIntv) nIntv = Tsim/τ = 80
Fixed: pt 0.744
Dynamic: pmin, pmax, Nmin, Nmax, p1, p2 0.5, 0.8, 4, 15, 0.5, 0.8
Fixed: pt,pcmin 1
120
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

Linear: pt,pcmin(k) pt,pcmin1=1 and pt,pcmin2=0.5

Parabolic: pt,pcmin pt,pcmin1=1 and pt,pcmin2=0.5
Exponential: pt,pcmin pt,pcmin1=1 and pt,pcmin2=0.5, β=1
The main outcomes of this scenario can be summarized as follows:
• For pure and fixed/dynamic probabilistic algorithms, RCH decreases as noise-level
increases, due to the high packet-loss introduced by the high noise-level; and no measure is
taken by the existing probabilistic algorithms to accommodate the negative effect of
increasing noise-level.
• The DNDP algorithm with any of the suggested distribution functions for pt,pcmin(k) presents
a better performance in terms of increasing RCH in presence of noise by effectively
adjusting pt based on both k and pc. The results obtained demonstrated that the algorithm
provides the highest RCH for various network noise-level, when compared with
fixed/dynamic probabilistic algorithms.
• It can be seen from Figure 6 that the DNDP algorithm almost produces the same RCH
using fixed, linear, and parabolic pt,pcmin(k) distribution, while the exponential distribution
produces a lower RCH, but still it is higher than what can be obtained using fixed/dynamic
probabilistic algorithms. However, enhancing RCH is paid by slight increase in RET as
shown in Figure 7.
• In a noiseless environment, the fixed pt is taken to be equal to pt,avg of the dynamic
approach, which is 0.744. In a noisy environment, the average pt (pt,avg) depends on the area
under the pt,pcmin(k) curve, and it can be seen in Figure 5 that the area under the curve can be
arranged from the largest to the smallest as follows: fixed, parabolic, linear, and finally
exponential. For the noiseless environment, pt,avg for the fixed and the dynamic
probabilistic algorithms are equal and always less than that for the noisy environment. This
comes in line with the results presented in Figure 8.
• The dynamic probabilistic algorithm and the DNDP algorithm with any pt,pcmin(k)
distribution produced the same performance in noiseless environment (pc=1). This is
because for pc=1, the noise-correction factor (α) (Eqn. 5) and the noise-dependent
retransmission probability (pt(pc)) (Eqn. 3) are equal to zero. Consequently pt(k, pc)=pt(k)
(Eqn. 7), which means all algorithms use the same distribution for pt (pt=pt(k)).

Figure 5. The variation of pt,pcmin(k) with k Figure 6. RCH against pc for various algorithms
121
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

Figure 7. RET against pc for various algorithms. Figure 8. pt,avg against pc for various algorithms.

7. CONCLUSIONS
The main conclusion of this work is that using a k-dependent pt,pcmin (pt,pcmin(k)) enhances the
performance of the DNDP algorithm, where the RCH can be improved against slight increases
in RET or maintaining the same RCH against significant reduction in RET. The DNDP
algorithm with linear pt,pcmin(k) provides the optimum performance when compared with other
distributions. The results also demonstrated that the RCH of the DNDP algorithm is close to the
RCH of pure flooding for various network noise levels at less cost.
It is highly recommended to investigate the effect of node densities and nodes speed and other
network parameters on the performance of the DNDP algorithm with various distribution for
pt,pcmin(k).

REFERENCES
[1] M. Bani-Yassein, M. Ould-Khaoua, L. Mackenzie, and S. Papanastasiou, “Performance
Analysis of Adjusted Probabilistic Broadcasting in Mobile Ad Hoc Networks, International
Journal of Wireless Information Networks, Vol. 13, No. 2, pp. 127-140, 2006.
[2] D. Scott and A. Yasinsac, “Dynamic Probabilistic Retransmission in Ad Hoc Networks”,
Proceedings of the International Conference on Wireless Networks (ICWN’04) (Editors: H.
R. Arabnia, L. T. Yang, and C. H. Yeh), Las Vegas, Nevada, USA, CSREA Press, Vol. 1,
pp. 158-164, 2004.
[3] E. Royer and C. Toh, “A Review of Current Routing Protocols for Ad Hoc Mobile
Wireless Networks”, IEEE Personal Communication Magazine, Vol. 6, No. 2, pp. 46-55,
1999.
[4] Z. Ismail and R. Hassan, “Evaluation of Ad Hoc on Demand Distance Vector Routing
Protocol in HetMAN Architecture”, Journal of Computer Science, Vol. 6, No. 7, pp. 830-
836, 2010.
[5] D. Johnson and D. Maltz, “Dynamic Source Routing in Ad Hoc Wireless Networks”,
Mobile Computing, Kluwer Academic Publishers, pp. 153-181, 1996.
[6] Z. Haas and M. Pearlman, “The Zone Routing Protocol (ZRP) for Ad Hoc Networks”,
Mobile Ad Hoc Network Working Group, IETF, 1998.

122
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

[7] Y. Ko and N. Vaidya, “Location-Aided Routing (LAR) in Mobile Ad Hoc Networks”,

Journal of Wireless Networks, Vol. 6, No. 4, pp. 307-321, 2000.
[8] A. Rahman, W. Olesinski, and P. Gburzynski, “Controlled Flooding in Wireless Ad Hoc
Networks. Proceedings of IEEE International Workshop on Wireless Ad Hoc Networks
(IWWAN’04), University of Oulu, Finland, 2004.
[9] M. Bani-Yassein and M. Ould-Khaoua, “Applications of Probabilistic Flooding in
MANETs”, International Journal of Ubiquitous Computing and Communication, Vol. 1,
No. 1, pp. 1-5, 2007.
[10] T. Tseng, S. Ni, Y. Chen, and J. Sheu, “The Broadcast Storm Problem in a Mobile Ad Hoc
Network”, Journal of Wireless Networks, Vol. 8, No. 2, pp. 153-167, 2002.
[11] H. Al-Bahadili, “Enhancing the Performance of Adjusted Probabilistic Broadcast in
MANETs. The Mediterranean Journal of Computers and Networks (MEDJCN), Vol. 6, No.
4, 2010.
[12] A. Qayyum, L. Viennot, and A. Laouiti, “Multipoint Relaying for Flooding Broadcast
Messages in Mobile Wireless Networks”, Proceedings of the 35th Hawaii International
Conference on System Sciences (HICSS’02), Island of Hawaii, pp. 3866- 3875, 2002.
[13] H. Al-Bahadili, R. Jaradat, “Performance Evaluation of an OMPR Algorithm for Route
Discovery in Noisy MANETs”, International Journal of Computer Networks and
Communications (IJCNC), Vol. 2, No. 1, pp. 85-96, 2010.
[14] C. Bettstetter, “The Cluster Density of a Distributed Clustering Algorithm in Ad Hoc
Networks”, Proceedings of the IEEE International Conference on Communications
(ICC’04), Paris, France, IEEE Computer Society Press, Vol. 7, pp. 4336-4340, 2004.
[15] H. Al-Bahadili and K. Kaabneh. Analyzing the Performance of Probabilistic Algorithm in
Noisy MANETs. The International Journal of Wireless & Mobile Networks (IJWMN), Vo.
2, No. 3, pp. 83-95, 2010.
[16] H. Al-Bahadili and Y Jaradat, “Development and Performance Analysis of a Probabilistic
Flooding in Noisy Mobile Ad Hoc Networks”, Proceedings of the 1st International
Conference on Digital Communications and Computer Applications (DCCA’07), Irbid,
Jordan, pp. 1306-1316, 2007.
[17] H. Al-Bahadili and A. Sabri, “A Novel Dynamic Noise-Dependent Probabilistic Algorithm
for Route Discovery in MANETs”, International Journal of Business Data Communication
and Networking (IJBDCN), Vol. 7, No. 1, pp. 52-67, 2011.
[18] H. Al-Bahadili, “On the Use of Discrete-Event Simulation in Computer Networks Analysis
and Design”, In Handbook of Research on Discrete-Event Simulation Environments:
Technologies and Applications (Eds.: E. Abu-Taieh and A. El-Sheikh,), Information
Science Reference, USA, pp. 414-442, 2009.
[19] Z. J. Haas, J. Y. Halpern, and L. Li, “Gossip-Based Ad Hoc Routing”, IEEE/ACM
Transactions on Networking, Vol. 14, No. 3, pp. 479-491, 2006.
[20] Y. Sasson, D. Cavin, and A. Schiper, “Probabilistic Broadcast for Flooding in Wireless
Mobile Ad Hoc Networks. Proceedings of IEEE Wireless Communications and
Networking (WCNC’03), Vol. 2, New Orleans, LA, USA, pp. 1124-1130, 2003.
[21] J. S. Kim, Q. Zhang, and D. P. Agrawal, “Probabilistic Broadcasting Based on Coverage
Area and Neighbor Confirmation in Mobile Ad Hoc Networks”, Proceedings of IEEE
Global Telecommunications Conference Workshops (GlobeCom’04), Dallas, Texas, USA,
IEEE Computer Society Press, pp. 96–101, 2004.

123
 International Journal of Wireless & Mobile Networks (IJWMN) Vol. 3, No. 2, April 2011

[22] C. Barrett, S. Eidenbenz, L. Kroc, M. Marathe, andL. Smith, ”Parametric Probabilistic

Routing in Sensor Networks”, Journal of Mobile Networks and Applications, Vol. 10, No.
4, pp. 529-544, 2005.
[23] K. Viswanath, and K. Obraczka, “Modeling the Performance of Flooding in Wireless
Multi-Hop Ad Hoc Networks”, Journal of Computer Communications, Vol. 29, No. 8, pp.
949-956, 2005.
[24] Q. Zhang and D. P. Agrawal, “Dynamic Probabilistic Broadcasting in MANETs”, Journal
of Parallel and Distributed Computing, Vo. 65, No. 2, pp. 220-233, 2005.
[25] J. Abdulai, M. Ould-Khaoua, L. Mackenzie, and M. Bani-Yassin, “On the Forwarding
Probability for On-Demand Probabilistic Route Discovery in MANETs”, Proceedings of
the 22nd Annual UK Performance Engineering Workshop (UKPEW’06), Poole, Dorset,
UK, pp. 9-15, 2006.
[26] J. Abdulai, M. Ould-Khaoua, and L. Mackenzie, “Improving Probabilistic Route Discovery
in Mobile Ad Hoc Networks”, Proceedings of the 32nd Conference on Local Computer
Networks (LCN ’07), Dublin, Ireland, IEEE Computer Society, pp. 739-746, 2007.
[27] I. Khan, A. Javaid, and H. Qian, “Coverage-Based Dynamically Adjusted Probabilistic
Forwarding for Wireless Mobile Ad Hoc Networks”, Proceedings of the 1st ACM
International Workshop on Heterogeneous Sensor and Actor Networks (HeterSanet ‘08),
Hong Kong, China, 2008, pp. 81-88, 2008.
[28] A. Hanash, A. Siddique, I. Awan, and M. Woodward, “Performance Evaluation of
Dynamic Probabilistic Broadcasting for Flooding in Mobile Ad Hoc Networks”, Journal of
Simulation Modeling Practice and Theory, Vol. 17, No. 2, pp. 364-375, 2009.
[29] H. Al-Bahadili, “Enhancing the Performance of Adjusted Probabilistic Broadcast in
MANETs”, The Mediterranean Journal of Computers and Networks (MEDJCN), Vol. 6,
No. 4, 2010.
AUTHORS
Hussein Al-Bahadili (Hbahadili@uop.edu.jo) received his B.Sc degree in Engineering from
University of Baghdad, Iraq, in 1986. He received the M.Sc and PhD degree from University
of London (Queen Mary College), UK, in 1988 and 1991, respectively. He is currently an
associate professor at Petra University, Jordan. He is also a visiting researcher at the Centre of
Wireless Networks and Communications (WNCC), University of Brunel (UK). He has
published many papers in leading journals and world-level scholarly conferences. He also
published a number of book chapters in high standard IT books. His research interests include computer
networks design and architecture, routing protocols optimizations, parallel and distributed computing,
cryptography and network security, and data compression.
Ghassan F. Issa (gissa@uop.edu.jo) received his B.E.T degree in Electronic Engineering
from the University of Toledo, Ohio, in 1983, and B.S.EE in Computer Engineering from Tri-
State University, Indiana in 1984. He received his M.S. and Ph.D. in Computer Science from
Old Dominion University, Virginia, in 1987 and 1992 respectively. He was a faculty member
and department chair of Computer Science at Pennsylvania College of Technology (Penn
State) from 1992–1995. He also served as faculty member and the dean of Computer Science
at the Applied Science University in Amman, Jordan from 1995-2007. Currently he is an associate professor
and the dean of Computer Science at Petra University, Jordan. His research interest covers block cipher, and
authentication.
Alia Sabri (a_sabri@asu.edu.jo) received her B.Sc degree in Computer Science from Jordan
University of Science and Technology (Jordan) in 2002. She received her M.Sc and PhD in
Computer Information Systems from the Arab Academy for Banking & Financial Sciences
(Jordan) in 2005 and 2009, respectively. She is currently working as an assistance professor at
the Applied Science University, Amman, Jordan. Her current research interests are in
developing efficient dynamic routing protocols for mobile ad hoc networks, wireless
networks management and security, and ad hoc networks modeling and simulation.
124