power aware ant colony routing algorithm for mobile ad-hoc ... › deanshipfiles ›...
TRANSCRIPT
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
ISSN: 1738-9984 IJSEIA
Copyright ⓒ 2015 SERSC
Power Aware Ant Colony Routing Algorithm for Mobile Ad-hoc
Networks
Alaa E. Abdallah1, Emad E. Abdallah
1, Feras Hanandeh
1, Ashraf Aljammal
1, and
Essam Al-Daoud2
1 Faculty of Prince Al-Hussein Bin Abdallah II For Information Technology,
Hashemite University, P.O. Box 150459, Zarqa 13115, Jordan
[e-mail: aabdallah,bsoul,emad,[email protected]] 2 Department of Computer Science, Zarqa University
Zarqa- Jordan
[e-mail: aabdallah, emad, feras, ashrafj @hu.edu.jo, [email protected] ]
Abstract
Due to the limited lifetime of nodes in ad hoc and sensor network, energy efficiency
needs to be an important design consideration in any routing algorithm. Most of the
existing Ant colony based routing algorithms grantee the packet delivery. However, they
suffer from the high power consumption due to the huge number of control messages to
establish and maintain a route from a source to a destination. This paper introduces two
new power-aware ant colony routing algorithms for mobile ad hoc network under three
main concurrent constraints. (1) Localized algorithms where only information about
neighbors' nodes is needed. (2) Maximize the algorithms delivery rate. (3) Minimize the
energy consumption. Our new algorithms are based on the idea of extracting a sub-graph
of the original network topology and combine it with the advantage of ant based routing
algorithms. Extensive experiments are conducted to prove that the new algorithms have
significant improvement on the network lifetime (up to twice) without affecting the high
delivery rate.
Keywords: Ad hoc network, Local algorithm, Energy efficient, Ant colony routing,
Geometric sub-graph
1. Introduction A Mobile ad hoc network (MANET) is a network formed of a set of wireless nodes
which communicate with other nodes without any supporting network infrastructure. In
MANET, a node can directly communicate with another node if and only if the two nodes
are located within the transmission range of each other. However, if a node wishes to
communicate with another node from outside the transmission range, it uses the help from
other nodes which play as a router to receive and forward packets. Mobile nodes in ad hoc
networks can move, join and leave the network freely. Thus, routing in such networks is a
crucial problem. However for many networks, a more important problem is providing an
energy efficient routing protocol because of the limited battery life of the wireless nodes.
Routing algorithms in mobile ad hoc network can be classified into two basic types [1] (1)
Proactive [2, 3], where each node in the network keeps routing information about every
other node up-to-date all the time. In this kind of protocol the nodes periodically
broadcast their routing tables to their neighbors. Proactive routing has a short response
time but it consume huge overhead exchanging routing information especially when the
mobility rate is big. (2) Reactive [4, 5, 6], where a node maintains only the routes that are
currently in use. In this strategy, if the route is not available, the current node holding the
packet issues a destination search request which is performed by flooding a short control
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
2 Copyright ⓒ 2015 SERSC
message. Obviously, there will be no routing table broadcast, thus less overhead but
longer time to find a route.
Another family of routing algorithms called position-based routing [1, 7, 8, 9, 10, 11].
Position-based routing algorithms assume that a node is aware of its position using cheap
instruments like GPS (Global Positioning System), the position of its neighbors by using
periodic hello messages, and the position of the destination using location service
algorithms [12, 13]. Position-based routing is known to have much less overhead than
proactive and reactive protocols but they may fail to find a route or they may find a non-
optimum route in some situations.
Over the last few years, researchers proposed a new set of routing algorithms that are
inspired by real ants’ behavior and their way to find the shortest path between food and
their nest [14, 15, 16, 17, 18]. The protocols works as follows [19, 20], while searching
for food, ants deposit on the ground a substance called pheromone to give them a way
back to the nest. By time the shorter paths would have more ants going through them, thus
more pheromone consecration. Other ants can smell pheromone, and when searching for a
path, they tend to choose, with high probability, paths marked by strong pheromone
concentrations. Most of ant-colony routing algorithms suffer from the high power
consumption due to the huge number of control messages which are needed to establish a
route from the source to the destination.
Many routing strategies use a sub-graph of the unit disk graph such that only the edges in
the sub-graph are used for routing. Therefore, much research effort has gone into the
development of algorithms for ad hoc network topology control (see [21, 22, 23] for
surveys). In this paper, we propose a new set of routing algorithms for mobile ad hoc
network which utilize the network nodes positions to extract a sub-graph of the original
Unit Disk Graph (UDG). Then we apply ANTHOCNET [24] routing over the extracted
sub-graph. Our simulations show that the new algorithms have significant improvement
on the network lifetime.
The rest of this paper is organized as follows. Section 2 is devoted to background
material. In Section 3, we briefly review some related position-based and ant-based
routing algorithms. In Section 4 we give detailed descriptions of the new proposed
routing algorithms. In Section 5, we present experimental results to demonstrate the
much improved performance of the proposed methods in comparison with existing
techniques. Finally, we conclude and point out some future directions in Section 6.
2. Preliminaries
2.1. The Network Model
We assume that a set of n wireless hosts is represented as a point set V in the 2D
space. All nodes have the same communication range of r, which represented as a
circle of radius r. Let ( ) be the Euclidean distance between the nodes u and
v: ( ) √( ) ( )
. Two nodes are connected by an edge if the
Euclidean distance between them is at most r. The resulting graph is called a unit
disk graph UDG(V, r). For node u, we denote the set of its neighbors by N(u). A
path from s to d is a sequence of nodes s = v1, v2, ….., vk = d, such that (vi ,vi+1) are
neighbors.
2.2 The Routing Problem
Given a unit disk graph UDG(V) corresponding to a set of points V , and a pair (s,d)
where , the problem of routing is to discover a path in UDG(V) from s to d. It’s
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
Copyright ⓒ 2015 SERSC 3
not an easy task to design a power efficient routing algorithm for many reasons (1) the
wireless links in the network are not known to any node in the network because nodes
may not have fixed locations. (2) Number of neighbors keeps changing because of the
interference and noise which affect the transmission range. (3) Nodes in the network can
only see neighbors and does not have a global look to the whole network.
We are interested in the following performance measures for routing algorithms: the
delivery rate [11] which is the percentage of times that the algorithm succeeds in
delivering its packet, the average packet delay, the overall overhead and the power
consumption [25, 26, 27, 28], it’s the amount of energy consumed by a message sent in
the network. If the transmission range is fixed for all the nodes, then the number of nodes
in the route path and the number of control messages are the main source of the energy
required for the routing task.
2.3 Related UDG sub-graphs
A distributed algorithm is called local if each node of the network only uses
information obtained uniquely from the nodes located no more than a constant
(independent of the size of the network) number of hops from it. Thus, during the
algorithm, no node is ever aware of the existence of other nodes in the network further
away than this constant number of hops. In the following we present some local
distributed algorithm to extract a sub-graph from UDG.
• Gabriel Graph (GG(G)) [29,30]: given any two adjacent nodes p and q in
UDG, the edge (p, q) belongs to GG if, and only if, the closed disc of which
line segment pq is a diameter contains no other nodes of V , see Figure 1. It
is known that if UDG is 2-dimensional then the Gabriel sub-graph is planar,
where the edges intersect only at their common end-vertices, and if the UDG
is connected, then its Gabriel sub-graph is also connected [31].
• YAO_K(G) Graph [32, 33]: with an integer parameter k ≥ 6 is defined as follows. Each node p divides the unit disk around it to k equal cones, see Figure 2. In each cone, only the edge (p,q) to the nearest neighbor q, if any, is part of the directed YAO_K(G). Ties are broken arbitrarily. The undirected graph obtained if the direction of each edge is ignored. If the UDG is connected, then its YAO_K(G) Graph is also connected [34, 35].
• Half Space Proximal Graph HSP(G) is defined as follows [36]. Each node p in UDG performs the following procedure: p draws an edge (p,q) to the nearest neighbor q, and draws a perpendicular line intersecting the edge (p, q) at its midway point, all nodes in the half plane contains q are discarded. The procedure then continues with the next non discarded node. See Figure
Figure 1. Edge x is counted in GG(G) graph if the disc of diameter x contains no other
nodes
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
4 Copyright ⓒ 2015 SERSC
3. It has been proved in [36] that HSP(G) is connected, has an out-degree of at most six, and contains the minimum weight spanning tree as its sub_graph.
3. Related routing algorithms
Routing in Mobile ad hoc network depends on many factors including network
topology, the type of information available during routing, and the specific underlying
network characteristics that could be used to define a heuristic to find a path quickly and
efficiently. The following reviews some representative ant-based and position-based
routing algorithms that are closely related to our proposed approach. We briefly discuss
the algorithmic methodologies as well as their limitations.
Greedy routing [30]: a position-based routing, where the current node c forwards
the packet to the neighbor node p that minimizes the remaining distance to the
destination node d. the same procedure is repeated until the destination node is
reached or no such node p exists. This routing method suffers from the so called
local minimum phenomenon, in which a packet may get stuck at a node that does
not have a neighbor that makes a progress to the destination, even though the
source and the destination are connected by a path in the network.
DREAM [37]: in this algorithm, the current node c forwards the packet to all
neighbors in the direction of the destination d. A node is considered to be in the
direction of d if it is located in a 2D cone that starts at current node and ends with
a circle centered at d, the circle radius equal to vmax * (t1 − t0) where t1 is the
current time, t0 is the time stamp of the position information that c has about d
and vmax is the maximum speed of the node in the network.
Figure 2. The shortest edge in each cone is counted in YAO_ K(G)
Figure 3. HSP(G) graph, the shortest edge is chosen, all other nodes behind the
perpendicular line are discarded
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
Copyright ⓒ 2015 SERSC 5
ANTHOCNET [24] is a reactive ant colony optimization algorithm. To find a
route from a source to a destination, two types of control packet are generated,
forward and backward ants. Over a specific time interval, the source node
generates periodic forward ants (control packet) for the destination. The ants
flood the network through all possible ways Figure 4. Each ant updates the weight
of every visited link by a fixed number (deposit on the ground a substance called
pheromone). Obviously ants that pick the shortest path from the source to the
destination will arrive first to the destination. At this point the destination will
generate backward ants to go through the exact same shortest path which will
change the status of the links by increasing the pheromone concentration on the
links along the path. The source node knows now the shortest path from the
source to the destination. To prevent infinite loops in the route discovery process,
each visited node u keeps the following about the forward ants: the identity of the
forward ants and the arrival time to u. When an ant visits node u, it checks the
ants arrival time and its identity. If the time are within a certain limit of those
produced by another ant of the same generation then the ant is forwarded.
Otherwise, the ant is discarded. The drawback of ANTHOCNET is the number of
forward and backward control messages that needs to be sent in the network for
establishing routes to the destination and the time needed before paths between
the nodes of the network is established.
ANTNET [38]: considered to be a proactive ant colony routing algorithm for
mobile switch networks. The algorithm starts by lunching a forward ant at regular
intervals from the source node. Depend on the information stored at the current
node routing table, the next node is selected as follows: suppose that a forward
ant is currently residing in node n and this node has k neighbors h1,h2, ...,hk, and
Φi is the amount of pheromone assigned to nhi. The forward ant will select hi as
the next node with a probability pi which is calculated using Equation 1. The
probability of selecting the next node is proportional to the pheromone
concentration. As in ANTHOCNET, when the forward ant reaches the
destination, it generates a backward ant that goes through the opposite direction
Figure 4. ANTHOCNET route discovery process, to find a route from S to D, S
generates periodic forward ants (control packet) for the destination D. The ants flood
the network through all possible ways.
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
6 Copyright ⓒ 2015 SERSC
of the path chosen by the corresponding forward ant. The backward ant updates
pheromone values as it moves on its way to the source node. Because ANTNET
route discover process select one neighbor at each intermediate node, the end to
end delay is higher than many other ant colony routing algorithms. In General
most of the ant colony routing algorithms including ANTHOCNET have a very
high delivery rate and find routes whose lengths are very close to the length of the
shortest path [39, 38].
4. Proposed Routing Algorithms
In this section we explain our new proposed algorithms, ANTSUB(HSP(G)) and
ANTSUB(YAO_ 6(G)). They are able to find nearly optimum routes and save up to 50%
of the energy consumption compared with the famous ANTHOCNET routing. As stated
before, ANTSUB(HSP(G)) and ANTSUB(YAO_6(G)) are reactive algorithms, thus a
route is searched only when a node wants to send data to another node outside its
transmission range. The route discovery is done only through control ant messages. Data
packets to be sent after the route discover process is finished and the route is established.
In order to decrease the energy consumption at each node in the network, we have to keep
the number of control messages as small as possible.
ANTSUB(HSP(G)): consider a source node, s, has a set of data packets and wants to send
them to a destination node, say d. Since s does not know how to reach d, it starts a route
discover process as follows: using HSP(G) algorithm s finds the set of edges around it that
belongs to HSP(G), then s sends several forward ants with unique sequence numbers
through each edge at regular time intervals, see Figure 5. Each neighbor receives a
forward ants run the HSP(G). Based on the values of pheromone trails the current node
selects the next hop as follows: consider the forward ant is currently at node n and this
node has k neighbors h1,h2, ...,hk, and Φi is the amount of pheromone assigned to nhi. The
forward ant will select hi as the next node with a probability pi which is calculated using
Equation 1. As on any ant colony routing algorithm, when an ant visits a node through an
edge, it deposits a specific amount of pheromone on that edge. The information about the
values of pheromone on the neighbors edges is stored on a table at each node (which is
going to be used while sending the data packets).
∑
(1)
In addition to the pheromone trial table, each node keeps another table which we call back
routing table. The table saves the id of each ant enters the node, the ant source node, the
ant destination node and the ant time stamp. This is used to discard duplicate ants. With
the help of the same table, the backward ant traverses in the reverse direction. If a node
does not hear any packet going to a given destination for a specific amount of time (order
of seconds), it deletes the back routing table. The pheromone trails for that destination
will be deleted from the pheromone trail table as well. Algorithm 1 shows in details the
route discovery process.
We have tried to use YAO_6(G) instead HSP(G) graph during the routing process of the
Algorithm 1 to a get a new routing algorithm ANTSUB(YAO_6(G)). It did not show
better power saving than ANTSUB(HSP(G)) but still it’s better than other ant colony
routing algorithms.
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
Copyright ⓒ 2015 SERSC 7
5. Simulation results
In this section we describe our simulation environment, evaluate the performance
of our new algorithms under several simulation scenarios and compare them with
other related routing algorithms.
5.1 Simulation Environment
The event-driven simulator can be described as follows:
1. We define 300mX300m square
2. A set V of n points (where n 100, 150, 200, 250, 300) is randomly
positioned in the square, this would lead to different node densities.
3. Each node has a transmission range equal to 15m.
4. If the network graph is not fully connected, all connected components in
the graph are calculated.
5. The largest connected component (LCC) among all the connected com-
ponents is selected to perform the routing algorithms.
6. For each network, one source-destination pair is selected randomly from
the LCC.
7. It is suggested in [24] to consider simulations with node density per unit
disk of around 5 in 2D environment. The proposed simulation environment
satisfies this density.
8. To estimate the performance, the algorithms are executed on 300 networks
and report the average results.
9. The algorithms are evaluated in terms of delivery rate, the average packet
delay, the overall overhead, and the power consumption of the whole
network.
10. The delivery rate is defined as the ratio of the number of packets received
by the destination node to the number of packets sent by the source node.
Figure 5. ANTSUB(HSP(G)) route discovery process, to find a route from S to D. Each
node receives a forward ant extracts the edges around C that belongs to HSP(G) and
broadcast the ant to all neighbors in HSP(G). This process is done in a periodic basses.
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
8 Copyright ⓒ 2015 SERSC
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
Copyright ⓒ 2015 SERSC 9
11. To calculate the average power consumption, the model proposed by
Heinzelman et al. [40] is utilized, which assumes the radio dissipates
Eelec=50nJ/bit power to run the transmitter. Thus both transmitter and
receiver nodes consume Eelec to transmit one bit receiver circuitry.
12. The radio dissipates Eamp=100pJ/bit/m2 to run the transmit amplifier,
assuming d2 energy loss due to channel transmission, where d is the
distance between nodes. This implies the sender consumes (Eamp * d2)
power to transmit one bit.
13. According to the above wireless model, transmitting k−bits message at
distance d the transmitter expends ETx(k, d) = Eelec * k + Eamp * k * d2. To
receive the k−bits, the receiver node expends ERx(k) = Eelec * k.
5.2 Observed Result
In Figure 6, we studied the effect of the running time on the average delivery rate of
the new proposed algorithms and compared them with other well-known routing
algorithms. Clearly, all studied routing algorithms reach 100% delivery rate by time,
except of greedy algorithm due to the local minimum problem. Our algorithms
ANTSUB(HSP(G)) and ANTSUB(YAO_6(G)) reaches 100% delivery rate way faster
than ANTNET. The reason is that the new algorithms generate discovery ants though all
possible ways at the source node and employ a connected sub-set of the original set of
edges used by ANTNET, which means less chance to go through wrong paths, thus the
shortest and the accurate path will be reached faster. ANTNET forwards ants to only one
neighbor depend on the information stored at the node itself, which may lead to a longer
path. Hence, more time to reach the destination.
Figure 6. The average delivery rate of (HSP(G)), ANTSUB(YAO_6(G)), and other routing
algorithms at different time clock. The number of nodes equal to 150
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
10 Copyright ⓒ 2015 SERSC
In Figure 7, we studied the effect of the node density on the delivery rate. The running
time is fixed in this study (80ms). The new proposed algorithms ANTSUB(HSP(G))
ANTSUB(YAO_6(G)) share highest delivery rate with ANTHOCNET. By increasing the
node density, greedy algorithm delivery rate increases. This can be explained by the
number of links, more links means more paths to the destination and less chances to face a
local minimum. The delivery rate of all examined ant colony routing algorithms decreases
by increasing the number of nodes because they all need more time to find an accurate
route to the destination. Since, ANTSUB(HSP(G)) uses less number of edges than
ANTSUB(YAO_6(G)) it has higher delivery rate.
Figure 7. The average delivery rate of (HSP(G)), ANTSUB(YAO_6(G)), and other routing
algorithms at different network density. The results are taken after 80ms.
Figure 8. The average packet delay of (HSP(G)), ANTSUB(YAO_6(G)), and other routing
algorithms at different time clock. The number of nodes equal to 150
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
Copyright ⓒ 2015 SERSC 11
In Figure 8 and Figure 9, we studied the delay of our proposed routing algorithms and
compare them with ANTNET and ANTHOCNET. In both figures we calculated end to
end delay based on the number of hops visited by each packet. Figure 8 shows the average
packet delay vs. time, in the beginning, the average packet delay of ANTHOCNET is
smaller than all other routing techniques. The average delay of ANTSUB(HSP(G)) and
ANTSUB(YAO_6(G)) reach to their minimum faster than the ANTNET. As time passes
on, all algorithms converge to the paths with almost the same lengths, where packets
experience almost the same delay. Figure 9 studied the effect of increasing number of
nodes against the delay; as expected increasing number of nodes will increase the delay
for all studied algorithms because the length of the path to the destination will increase.
Figure 9. The average packet delay of (HSP(G)), ANTSUB(YAO_6(G)), and other routing
algorithms at different network density. The results are taken after 80ms.
Figure 10. The average power consumption of (HSP(G)), ANTSUB(YAO_6(G)), and other
routing algorithms at different time clock. The number of nodes equal to 150
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
12 Copyright ⓒ 2015 SERSC
The main contribution of the paper can be found in Figure 10, where we show the average
power consumption from the whole network after running the algorithm for a specific
amount of time; clearly the ANTHOCNET has consumed huge power to discover and
maintain routes. It’s almost double the power consumed by our proposed algorithms.
ANTNET has the lowest power consumptions compared with other studied algorithms;
this due to the nature of the algorithm, each node receives an ant, instead of re- broadcast
it, the ant is forwarded to one of the neighbors. ANTSUB(HSP(G)) and
ANTSUB(YAO_6(G)) send ants to all source neighbors in the sub-graph and then uses
similar forwarding schema to ANTNET. The power consumption is close to ANTNET
since they use a sub-graph of UDG which lead to reduce the number of ants, because the
pheromone concentration will increase on the right path much faster. After the
millisecond 160, in all algorithms the power consumption stops increasing dramatically
because most of the algorithms usually finds route by this time, thus they stop generating
discovery ants. Figure 11, depict how the number of nodes affect the overall power
consumption of all algorithms. As predicted, ANTSUB(HSP(G)), ANTSUB(YAO_6(G)),
and ANTNET have much less power consumption than ANTHOCNET.
It is essential to keep the overhead of any ad-hoc routing algorithm as small as possible.
Large overhead makes a routing algorithm not scalable and sometimes useless. One of the
necessary parameters that should be kept small is the amount of control traffic exchanged
in the network when a routing protocol is running. Figure 12 displays the number of
generated ants by the new proposed algorithms along with the other studied algorithms in
different clock times. Clearly, the number of generated ant in ANTHOCNET grows very
quickly specially before the millisecond 160. For ANNET the number of ants grows in a
linear function of time, because ANTNET generates an ant at every time clock. The total
number of ants generated in ANTSUB(HSP(G)) and ANTSUB(YAO_6(G)) is relatively
higher than ANTNET but it still raises linearly, because each node has a bounded degree
after extracting the sub-graph.
Figure 11. The average power consumption of (HSP(G)), ANTSUB(YAO_6(G)), and other
routing algorithms at different network density. The results are taken after 80ms.
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
Copyright ⓒ 2015 SERSC 13
6. Conclusions
In this paper, we propose two new ant colony routing algorithms for mobile ad hoc
network (ANTSUB(YAO_6(G)), and ANTSUB(HSP(G))). We utilize the network
nodes positions to extract a sub-graph of the original UDG graph. The route discovery
process of the new algorithms is done only over the extracted sub-graph. Simulation
results demonstrate that the new proposed algorithms have improved the network
lifetime dramatically, up to twice the lifetime of other ant colony routing algorithms,
without affecting the high delivery rate advantage. For future work, we plan to analyze
the relationship between number of nodes, size of sub-graph, transmission range,
delivery rate, and power consumption for further improvement.
References
[1] M. Mauve, J. Widmer, and H. Hartenstein. A survey of position-based routing in mobile ad-
hoc networks. IEEE Network Magazine, 15(6):30–39, 2001.
[2] C. Perkins and P. Bhagwat. Highly dynamic destination sequenced distance- vector routing
(DSDV) for mobile computers. In Proc. of the SIGCOMM, Conference on Communications
Architectures, Protocols and Applications, pages 234–244, London, September 1994.
[3] T. Clausen and P. Jacquet. Optimized link state routing protocol (OLSR). In RFC 3626,
IETF Network Working Group, October 2003.
[4] D. Johnson and D. Malts. Dynamic source routing in ad hoc wireless networks. In Mobile
Computing (ed. T. Imielinski and H. Korth), chapter 5, pages 153–181. Kluwer Academic
Publishers, 1996.
[5] C. Perkins and E. Royer. Ad hoc on-demand distance vector routing. In Proc. of the 2nd
IEEE Workshop on Mobile Computing System and Application(WMCSA), pages 90–100,
San Juan, February 1999.
[6] S. Basagni, I. Chlamtac, and V. Syrotiuk. Dynamic source routing for ad hoc networks using
the global positioning system. In Proc. of the IEEE Wireless Communications and
Networking Conference(WCNC’99), pages 21–24, New Orleans LA, September 1999.
[7] A.E. Abdallah, T. Fevens, J. Opatrny, and I. Stojmenovic. Power-aware 3d position-based
routing algorithms using adjustable transmission ranges for ad hoc and sensor networks. Ad
Hoc Networks, 8(1):15–29, 2010.
[8] A.E. Abdallah, T. Fevens, and J. Opatrny. High delivery rate routing algorithms for 3d ad
hoc network. Computer Communications, 31(4):807– 817, 2008.
[9] S. Giordano, I. Stojmenovic, and L. Blazevic. Postion based routing algorithms for ad hoc
Figure 12. The average number of generated ants of (HSP(G)), ANTSUB(YAO_6(G)), and
other routing algorithms at different time clock. The number of nodes equal to 150
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
14 Copyright ⓒ 2015 SERSC
networks for ad hoc networks: A taxonomy. In Ad hoc wireless Networking (ed. X. Cheng,
X. Huang, and D.Z. Du), Kluwer, pages 103–136, July 2003.
[10] P. Bose, P. Morin, I. Stojmenovic, and J. Urrutia. Routing with guaranteed delivery in ad
hoc wireless networks. Wireless Networks journal, 7(6):609– 616, 2001.
[11] T. Fevens, A.E. Abdallah, and B. Bennani. Randomized ab-face-ab routing algorithms in
mobile ad hoc network. In 4th International Conference on AD-HOC Networks and
Wireless, volume 3738 of LNCS, pages 43–56, Mexico, October 2005.
[12] J. Hightower and Borriello G. Location systems for ubiquitous computing. Computer,
34(8):57–66, 2001.
[13] J. Li, J. Jannotti, D. De Couto, D. Karger, and R. Morris. A scalable location service for
geographic ad-hoc routing. In Proc. of the 6th ACM In- ternational Conference on Mobile
Computing and Networking (MobiCom), pages 120–130, Boston, August 2000.
[14] S. Kamali and J. Opatrny. A position based ant colony routing algorithm for mobile ad-hoc
networks. Jornal of Networks, 3(4):31–41, 2008.
[15] Caro Gianni Gambardella Ducatelle, Frederick and L. Maria. Ant Agents for Hybrid
Multipath Routing in Mobile Ad Hoc Networks. In Proceedings of the conference on
Wireless On-demand Network Systems and Services, pages 44–53, Swizerland, January
2005.
[16] P. Thulasiraman J. Wang, E. Osagie and R. Thulasiram. Hopnet: A hybrid ant colony
optimization routing algorithm for mobile ad hoc network. Ad Hoc Networks, 7(4):690 –
705, 2009.
[17] G. Caro M. Saleem and M. Farooq. Swarm intelligence based routing proto- col for wireless
sensor networks: Survey and future directions. Information Sciences, 181(20):4597 – 4624,
2011.
[18] S. Xu M. Liu and S. Sun. An agent-assisted qos-based routing algorithm for wireless sensor
networks. Journal of Network and Computer Applications, 35(1):29 – 36, 2012.
[19] D. Caro Gianni Gambardella Dorigo, Marco and L. Maria. Ant algorithms for discrete
optimization. Artif. Life, 1999.
[20] Li. Ang A. Zungeru and K. Seng. Classical and swarm intelligence based routing protocols
for wireless sensor networks: A survey and comparison. Journal of Network and Computer
Applications, 35(5):1508 – 1536, 2012.
[21] Yun Zhao. Motion vector routing protocol: A position based routing protocol for mobile ad
hoc networks. In PhD. thesis, University Of Arizona, USA, April 2005.
[22] Xiang-Yang Li. Topology control in wireless ad hoc networks. Book Chapter of Ad Hoc
Networking, IEEE Press, edited by Stefano Basagni, Marco Conti, Silvia Giordano, and Ivan
Stojmenovic., pages 175–203, 2003.
[23] R. Rajaraman. Topology control and routing in ad hoc networks: a survey. ACM SIGACT
News, 33(2):60–73, 2002.
[24] F. Ducatelle G. Caro and L. Gambardella. Anthocnet: An adaptive nature- inspired
algorithm for routing in mobile ad hoc networks. Self-Organisation in Mobile Networking,,
16(5):443–455, 2005.
[25] G. Karmakar Kh. Alam, J. Kamruzzaman and M. Murshed. Dynamic adjustment of sensing
range for event coverage in wireless sensor networks. Journal of Network and Computer
Applications, 46(0):139–153, 2014.
[26] I. Stojmenovic and X. Lin. Power aware localized routing in ad hoc networks. IEEE Trans.
on Parallel and Distributed Systems, 12(11):1122– 1133, 2001.
[27] C. Yu, B. Lee, and H. Youn. Energy efficient routing protocols for mobile ad hoc networks.
Wireless Communications and Mobile Computing Journal, 3(11):959–973, 2003.
[28] B. Zhang and H. Mouftah. Energy-aware on-demand routing protocols for wireless ad hoc
networks. Wireless Networks, 12(4):481–494, 2006.
[29] K. Gabriel and R. Sokal. A new statistical approach to geographic variation analysis.
Systematic Zoology, 18(3):259–278, 1969.
[30] G. Kao, T. Fevens, and J. Opatrny. Position-based routing on 3D geometric graphs in mobile
ad hoc networks. In Proc. of the 17th Canadian Conference on Computational Geometry
(CCCG’05), pages 88–91, Windsor, August 2005.
[31] P. Bose, L. Devroye, W. Evans, and D. Kirkpatrick. On the spanning ratio of gabriel graphs
and beta-skeletons. In Proc. of the Latin American Theoretical Infocomatics (LATIN), pages
479–493, Mexico, April 2002.
[32] P. Bose, J. Gudmundsson, and P. Morin. Ordered theta graphs. Computa- tional Geometry:
Theory and Applications, 28(1):11–18, 2004.
[33] A.C. Yao. On constructing minimum spanning trees in k-dimensional spaces and related
problems. SIAM Journal on Computing, 11(4):721–736, 1982.
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
Copyright ⓒ 2015 SERSC 15
[34] J. Keil and C. Gutwin. Classes of graphs which approximate the complete euclidean graph.
Discrete Computational Geometry, 7(1):13–28, 1992.
[35] X. Li, P. Wan, and Y. Wang. Power efficient and sparse spanner for wireless ad hoc
networks. In Proc. of the IEEE International Conference on Computer Communications and
Networks (ICCCN01), pages 564–567, Arizona, October 2001.
[36] E. Chaves, S. Dobrev, E. Kranakis, J. Opatrny, L. Stacho, H. Tejeda, and J. Urrutia. Half-
space proximal: A new local test for extractinga bounded dialation spanner of a unit disk
graph. In Proc. of the 9th International Conference on Principles of Distributed Systems
(OPIDIS2005), pages 235–245, Italy, December 2006.
[37] S. Basagni, I. Chlamtac, V. Syrotiuk, and B. Woodward. A distance routing effect algorithm
for mobility (DREAM). In Proc. of the 4th annual ACM/IEEE international conference on
Mobile computing and networking (MOBICOM), pages 76–84, Dallas, 1998.
[38] U. Sorges M. Gunes and I. Bouazizi. Ara - the ant-colony based routing algorithm for
manets. In Proceedings of the International Conference on Parallel Processing Workshops,
pages 79–89, Vancouver, August 2002.
[39] G. Caro and M. Dorigo. Ant colonies for adaptive routing in packet- switched
communications networks. In Proceedings of the 5th ACM In- ternational Conference on
Parallel Problem Solving from Nature, pages 673–682, 1998.
[40] W. Heinzelman, A. Chandrakasan, and H. Balakrishnan. Energy-efficient communication
protocols for wireless microsensor networks. In Proc. of the International conference on
System Sciences, pages 3005–3014, Hawaii, January 2000.
Authors
A.E. Abdallah, is currently an Assistant Professor in the Department of
Computer Science at the Hashemite University (HU), Jordan. He received his
PhD in Computer Science from Concordia University in 2008, where he
worked on routing algorithms for mobile ad hoc networks. He received his
BS from Yarmouk University, Jordan, and MS in from the University of
Jordan in 2000 and 2004, respectively. Prior to joining HU, he was a network
researcher at consulting private company in Montreal (2008 - 2011). His
current research interests include Routing Protocols for Ad Hoc Networks,
Parallel and Distributed Systems, and Multimedia Security.
E.E. Abdallah, is currently an Associate Professor in the Department of
Computer Information Systems at the Hashemite University (HU), Jordan.
He received his PhD in Computer Science from Concordia University in
2008, where he worked on multimedia security, pattern recognition and 3D
object recognition. He received his BS in Computer Science from Yarmouk
University, Jordan, and MS in Computer Science from the University of
Jordan in 2000 and 2004, respectively. Prior to joining HU, he was a
Software Developer at SAP Labs Montreal.
Feras Hanandeh, is currently an Associate professor at the Prince Al
Hussein Bin Abdullah II Faculty of Information Technology, Hashemite
University, Jordan. He obtained his PhD in computer science from
University Putra Malaysia, Malaysia in 2006. His current research in terests
include distributed databases, parallel databases focusing on issues related to
integrity maintenance, transaction processing, query processing and
optimization; grid computing, artificial intelligence and geographical
information systems
International Journal of Software Engineering and Its Applications
Vol. 9, No. 12 (2015)
16 Copyright ⓒ 2015 SERSC
Ashraf Aljammal, received the BSc degree in computer science from
Albalqa' Applied University, Al-Salt, Jordan, in 2006, the master's degree
from Science University of Malaysia, USM, Malaysia, in 2007, and the PhD
degree from Science University of Malaysia, USM, Malaysian, in 2011. He is
currently an assistant professor in the department of Computer Science of the
Hashemite University, Zarqa, Jordan, since 2012. His research interests
include network monitoring, network security, cloud computing, parallel and
distributed system security.
Essam Al-Daoud, is currently a Professor in the Department of
Computer Science at Zarka University. He received his BSc from Mu’tah
University, MSc from Al al-Bayt University and his PhD in Computer Science
from University Putra Malaysia in 2002. His research interests include
machine learning, soft-computing, cryptography, bioinformatics and quantum
cryptography.
.