8th Annual Communication Networks and Services Research Conference
An Energy-efficient Broadcast Protocol in MANETs
Design and Evaluation
Xiaoying Zhang, Thomas Kunz
Systems and Computer Communications Research Centre
Engineering 3701 Carling Avenue
Carleton University Ottawa, Ont., Canada K2H 8S2
Ottawa, Ont., Canada K1S 5B6 li.li@crc.ca
xiaoying0113@hotmail.com,
tkunz@sce.carleton.ca
Abstract—Energy efficiency is a critical issue in mobile ad hoc
networks (MANETs) for increasing the lifetime of the
individual nodes as well as the overall network. In this paper,
we propose a new energy-efficient broadcast protocol called
EBOLSR, which adapts the energy-efficient OLSR (EOLSR)
protocol to the broadcasting domain. In order to maximize the
broadcasting network lifetime, we take neighboring nodes’
residual energy into consideration when selecting multipoint
relay (MPR) nodes. We evaluate the performance of EBOLSR
by using NS2 and compare it with three other broadcast
protocols, Classical Flooding, Simplified Multicast Forwarding
(SMF), and the Coding-based Broadcast (CodeBCast)
protocol. Simulation results show that EBOLSR results in less
energy consumption and longer network lifetime than Classical
Flooding. We also explain why EBOLSR does not outperform
SMF or CodeBCast in terms of energy consumption and
network lifetime.
Keywords-MANETs; routing; broadcast; energy-efficieny;
OLSR; SMF
I. INTRODUCTION
A mobile ad hoc network (MANET) is a collection of
wireless mobile nodes dynamically forming a temporary
network without the use of any existing network
infrastructure or centralized administration. Due to the
limited transmission range of wireless network interfaces, a
packet needs to traverse multiple hops to a destination
beyond the range of its radio signal. Without a stationary
infrastructure such as base stations, each node operates not
only as a host but also as a router. Considering the energy
and bandwidth constraint experienced by a typical MANET
node powered by batteries, a practical routing protocol
should minimize resource consumption and meanwhile
maximize network efficiency.
Broadcasting a packet to the entire network is a basic
operation and has extensive applications in MANETs. Many
unicast routing protocols, such as the Dynamic Source
Routing (DSR) [1] protocol and the Ad Hoc On-demand
Distance Vector (AODV) [2] protocol, use broadcast to
discover unicast routes toward destinations. In addition,
depending on the application, broadcast can be used to
deliver data packets to all nodes in the network. For
example, it may become necessary to have reliable
communications during large-scale emergency situations.
978-0-7695-4041-2/10 $26.00 © 2010 IEEE
DOI 10.1109/CNSR.2010.32
Li Li Oliver Yang
School of Information Technology
and Engineering
University of Ottawa
Ottawa, Ont., Canada K1N 6N5
yang@site.uottawa.ca
Information such as text, audio, or video may be broadcast or
multicast to survivors to inform them of shelter locations,
details of the disaster, how to respond to the disaster, and so
on. Moreover, situational awareness data may be broadcast
or multicast to various rescue teams such as the crisis center,
the police department, emergency medical services (i.e.,
hospitals, ambulances), and the fire department. In order to
maximize the broadcasting network lifetime, we propose a
new energy-efficient broadcast protocol by adpting the
energy-efficient selection strategy of multipoint relaying in
EOLSR [3].
In this work, we propose a new energy-efficient
broadcast protocol and then compare it with three other
broadcast protocols in MANETs. These three protocols are
Classical Flooding, Simplified Multicast Forwarding (SMF)
[8], and a coding-based broadcast protocol (CodeBCast) [9].
Classical Flooding is the simplest broadcast method, in
which each node retransmits packets to all its neighbors the
first time it receives it. It is reliable for MANETs with a low
density of nodes and high mobility, however, it may cause
severe network congestion and quickly exhaust the nodes’
energy, particularly in dense networks. SMF provides basic
IP multicast forwarding for MANETs, based on a duplicate
packet detection mechanisms for forwarding IP multicast
packets and efficient relay set mechanisms for reducing
contention and congestion in wireless multi-hop scenarios.
CodeBCast applies the network coding approach, which is
designed to achieve the maximum network throughput, to a
deterministic broadcast method, resulting in a significant
reduction in the number of transmissions in the network. The
proposed new energy-efficient broadcast protocol, EBOLSR,
extends the Energy-efficient OLSR (EOLSR) [3] protocol to
a broadcast protocol. In order to test and compare the
performance of the above protocols, we simulate two distinct
scenarios, a WiFi scenario and a Rescue scenario in NS2.
Not surprisingly, as the simulation results confirm, Classical
Flooding has the highest packet delivery ratio and energy
consumption among all, and CodeBCast has the lowest
energy consumption and longest average end-to-end delay.
However, some results are unexpected. Although we take
energy efficiency into account, EBOLSR, while achieving
results comparable to SMF in terms of the packet delivery
ratio and average end-to-end delay, does not outperform
SMF in terms of energy consumption and network lifetime.
Analyzing additional simulation results, we determine that
199
EBOLSR does not minimize energy consumption nor
maximize the network lifetime because more multipoint
relaying nodes are selected to rebroadcast in the EBOLSR
protocol than in SMF. In addition, the bigger size of Hello
packets in the EBOLSR protocol also causes it to consume
more energy than SMF.
The rest of this paper is organized as follows: Section 2
reviews the related energy-efficient broadcast protocols.
Section 3 presents how to extend EOLSR to EBOLSR in the
broadcast domain. Performance evaluations of the EBOLSR
protocol and comparisons with Classical Flooding, SMF [4]
and CodeBCast [5] protocols are given in Section 4. Section
5 concludes the paper.
II. RELATED WORK
A straightforward approach to broadcasting is Classical
Flooding, in which each node rebroadcasts a packet
whenever it is received for the first time. Although this
method is very reliable for MANETs that have low node
density and high network mobility, it may cause severe
network congestion, which is often referred to as the
broadcast storm problem [6]. The broadcast storm can
quickly exhaust the nodes’ energy.
In MANET, there are mainly two fundamental classes
of energy-aware broadcast problems that have been heavily
investigated recently [7]: the minimum energy broadcast
(MEB) problem and the maximum lifetime broadcast (MLB)
problem. The MEB problem aims to minimize the total
transmission power consumption of all nodes involved in the
broadcast session, solutions of which include the minimum
spanning tree (MST) [8], shortest path tree (SPT) [8],
broadcast incremental power (BIP) [8], etc. The MLB
problem aims to maximize the operation time until the
battery depletion of the first node involved in the broadcast
session. The minimum decremental lifetime (MDLT)
scheme [9], directed minimum spanning tree (DMST) [10],
and directed prim broadcast tree (DPBT) algorithm [11] are a
few of the proposed solutions for the MLB problem. Since
broadcasting can be considered as a special case of
multicasting, many of the existing maximum lifetime (e.g.
[11]) and minimum energy (e.g. [20]) multicast algorithms
can potentially be extended for broadcasting purposes.
Two approaches are mainly employed to develop energy-
efficient broadcast protocols. In the first approach, each node
can adjust its transmission power based on the distance to the
receiving node(s) and the background noise either
continuously or in a discrete fashion. Most of the previous
studies on energy-efficient broadcast focused on configuring
power levels of each node [12, 13, 14, 15, 16]. In the other
approach, all nodes in the network use a fixed power level
for transmissions with a goal to reduce the overall number of
retransmissions for broadcasting a packet. As a result, the
careful selection of retransmission nodes would reduce the
overall energy consumption in the broadcast. In this paper,
we propose our protocol based on the latter approach.
200
III. ENERGY-EFFICIENT BROADCAST OLSR
The OLSR routing protocol has been standardized by
the IETF [17]. However, this protocol does not take energy
consumption into account. [3] proposed a new energy-
efficient unicast routing protocol, EOLSR, which extends
OLSR in order to make it energy-efficient. EOLSR is
designed to maximize the network lifetime by selecting the
path with the minimum cost, where the cost includes the
residual energy of each visited node and the energy
consumption of a packet on this path. Based on the
multipoint relaying strategy in EOLSR, we propose a new
energy-efficient broadcast protocol, EBOLSR, which aims at
increasing the network lifetime for broadcast
communications.
A. Energy Consumption Model
A key consideration for any energy-efficient protocol is
the energy consumption at a wireless node. With respect to
network activities (i.e., ignoring energy consumed in lighting
up a display, power a hard drive, etc.), each node’s radio can
be in one of the following three states:
• Transmitting: the node is transmitting messages
with transmission power Pt,
• Receiving: the node is receiving messages with
reception power Pr,
• Idle: when no message is being transmitted, the
node stays idle and keeps listening to the medium,
consuming energy at a rate that corresponds to a
powerlevel Pidle.
In general, the following power inequalities hold (i.e.,
sending is more expensive than receiving, which in turn is
more expensive than operating in idle mode):
Pidle< Pr< Pt.
We assume that the power level of each state is fixed for
all nodes in the network. When a transmitter transmits one
packet, because of the shared nature of wireless medium, all
its neighbors receive this packet even if it is intended to only
one of them.
B. Energy-Efficient Selection of MPRs
Different from the MPR selection in OLSR, the energy-
efficient selection of MPRs is based on the residual energy in
the nodes. The selected nodes are thus denoted as EMPRs as
in EOLSR [3]. In EOLSR, three variants of EMPR selection
are discussed, applying different rules in computing the
residual energy, where ER(M) denotes the residual energy of
node M:
a) The E strategy considers only ER(M), the residual
energy of the EMPR candidate, M;
b) The M1E strategy considers the weighted residual
energy of the EMPR candidate M and its 1-hop
neighbors:
ER (M ) E ( D)
min( , min D∈1hop ( M ) ( R ))
Ptrans + Prcv 2 × p rcv .
 The weights of ER(M) and ER(D) take into account the IV. PERFORMANCE EVALUATION
role played by the nodes M and D in a retransmission In this section, we evaluate the performance of
from N, the node performing the EMPR selection, to Classical Flooding, SMF [4], CodeBCast [5] and EBOLSR.
D, one of its 2-hop neighbors, via the node M. It All protocols are implemented in NS2.29. Implementing
represents the maximum transmission duration that can Classical Flooding is relatively simple and only requires
be sustained; duplicate packet detection. Each packet is uniquely identified
c) The M2E strategy considers the weighted residual by the address of the sender and a sender-assigned ID. The
energy of the EMPR candidate M and its 1-hop and 2- version of SMF we used is the version provided by NRL
hop neighbors: [18], which is only supported for NS2.29. The CodeBCast
min(
ER (M ) E ( D)
, min D∈1hop ( M ) ( R
E ( D)
), min D∈2 hop ( M ) ( R )) protocol has already been implemented in [19]. A Reed-
Ptrans + Prcv 2 × prcv Prcv Solomon coding algorithm is used on top of SMF
Similar to M1E, the strategy takes into account the broadcasting. EBOLSR is implemented as a modification of
energy consequences of selecting a specific node as SMF.
MPR: that node will receive and rebroadcast a packet,
A. Network Scenarios and Performance Metrics
its one-hop neighbors will receive the packet twice,
and the two-hop neighbors at least once. To evaluate these four protocols, we simulated two
distinct types of networks: a WiFi Scenario and a Rescue
[3] evaluated the average number of EMPRs per node as Scenario with a more realistic radio link model based on
a function of network density with these three selection Ricean fading (called RiceanRescue) and low-bandwidth
variants, and compared it with the average number of MPRs tactical radios.
per node in OLSR. Simulation results showed that the
number of EMPRs per node is higher than the number of WiFi Scenario: The WiFi scenario is a scenario
MPRs per node. Furthermore, the M1E and M2E selection commonly used in MANET protocol evaluations. Simulation
strategies give comparable results and both outperform the E parameters are shown in Table I. For the SMF, CodeBCast
selection strategy. Since M1E is less complicated to compute and EBOLSR protocols, periodic Hello packets are
and needs less information from the network than M2E, we transmitted by each node every 1 second and the Hello jitter
selected the M1E selection strategy in our proposed protocol is 0.1 second. The jitter interval randomizes the Hello
to determine EMPRs. transmission time, this helps to avoid protocol
After exchanging Hello messages periodically, each synchronization issues and packet collision. All nodes in the
node has learned about its 1-hop and 2-hop neighbor sets. network have the same fixed transmission, reception and idle
The EMPR selection algorithm by any node, N, is described power. The initial energy of 25 nodes is uniformly
as follows: distributed in [100J-800J].
a) Calculate the maximum transmission duration that can
TABLE I. SIMULATION PARAMETERS FOR WIFI SCENARIO
be sustained,
ER (M ) E ( D) Simulation Parameters Value
min( , min D∈1hop ( M ) ( R ))
Ptrans + Prcv 2 × p rcv Number of Nodes 25
where ER(M) is the residual energy of N’s 1-hop Network Size 700m×700m
neighbor, and ER(D) is the residual energy of N’s 2- Simulation Time 600s
hop neighbor. Propagation Model Two-ray Ground
b) Calculate D(M), the degree of 1-hop neighbors of M, Medium Access Control IEEE 802.11
excluding the 1-hop neighbors of N. Transmission Range 250m
c) Sort the 1-hop neighbors in decreasing order of the Transmission Power 1.4W
selection criterion in step a). Receiving Power 1.0W
d) Select the first 1-hop neighbor, add it to the EMPR set, Idle Power 0.5W
and then remove its 1-hop neighbors in N’s 2-hop Bandwidth 2Mb/s
neighbor set. Packet Type CBR
e) Select the next 1-hop neighbor in order and repeat step Packet Size 512bytes
d), until all nodes in N’s 2-hop neighbor set are Packet rate 4Pkts/s
removed. If there is more than one node with the same Mobility Model Random Way-Point
maximum transmission duration, select the node with Maximum Speed 1m/s, 5m/s, 10m/s,
the highest D(M). 15m/s, 20m/s
d) The EMPR set is generated. If the 1-hop neighbors of Number of Senders 1
any node in the EMPR set are a subset of other nodes’ Number of Receivers 25
1-hop neighbors currently in the EMPR set, remove
this node from the EMPR set.

201
 RiceanRescue Scenario: In the RiceanRescue scenario, WiFi Scenario: Figure 1 shows the packet delivery ratio
the PHY model combines Ricean Fading and Shadowing, (PDR) as the function of mobility for the four protocols. For
with parameters based on field data from a CRC [19] all protocols, increasing the mobility of nodes leads to a drop
measurement study. The parameters are chosen such that a in PDR except for Classical Flooding. Classical Flooding has
link has near perfect reception up to 20 km, with packet loss the highest PDR in all mobility modes among all protocols
increasing thereafter. However, even for a transmission due to the redundant nature of this protocol. Since some
range of 40 km, a nontrivial amount of packets are still nodes fail during the simulation, Classical Flooding cannot
received (approximately 50%, given the Ricean K factor and achieve 100% PDR. CodeBCast has the second highest PDR.
path loss exponent obtained from the measurement data). As SMF and EBOLSR achieve almost the same results (and
the radio range increases, the network topology changes less therefore the curves overlap).
often. So the Hello interval is set to 10 seconds to minimize
the protocol overhead. The nodes all move with a velocity TABLE II. SIMULATION PARAMETERS FOR RICEANRESCUE SCENARIO
that depends on the group a node belongs to. Of the 50
nodes, 3 nodes, representing command-and-control centers, Number of Nodes 50
are nearly stationary. 7 nodes move individually around the Network Size 40000m*40000m
whole simulation area based on the Random Waypoint Simulation Time 2000s
mobility model, with speed randomly selected between 30 Propagation Model RiceanShadowing
and 70 km/h and 0 pause time. The remaining 40 nodes are Ricean K Factor 4.8dB
grouped into 4 sets of ten nodes each, moving as a group. Reference Distance d0=100m
Each group of 10 nodes moves according to the Reference Path Loss Exponent 3.75
Point Group Mobility Model, where the reference point Shadowing Loss 4.5
moves with a speed randomly selected between 30 and 70 Variance
km/h and 0 pause time. Within each group, nodes can Transmit Power 0.281838W
deviate from the reference point by +/- 1 km in each
Frequency 7.75e+07
direction. In addition, each of the four groups is assigned to
work in one quadrant of the simulation area, with the Receiving threshold 6.61502e-16
quadrants slightly overlapping. In all scenarios, the network Medium Access Control TDMA
is and remains fully connected. In this network scenario, we Initial Energy 1000J
assume that nodes are equipped with tactical radios, Transmission Power 1.4W
operating at 128 kbps. The application traffic mimics a VoIP Receiving Power 1.0W
traffic flow from a single source to all nodes in the network. Idle Power 0.5W
The simulation parameters for this scenario are summarized Bandwidth 128Kb
in Table II. Packet Type CBR
B. Performance Comparison Packet Size 80bytes
Data rate 2.4Kbps
The performances of the four protocols are compared
Number of Senders 1
below. All results are averaged over 10 simulation runs.
Number of Receivers 50

Figure 1. Packet delivery ratio

202
 Figure 2. Average end-to end delay

Figure 2 shows the average end-to-end delay as the
function of the mobility speed. Since all nodes in Classical
Flooding rebroadcast the received packet if it is not
duplicated, resulting in network congestion, Classical
Flooding starts to experience increased packet delays
compared to SMF and EBOLSR, which have the lowest
delay (again, nearly identical). CodeBCast, not
surprisingly, has the highest delay because some packets
need to be buffered in the sending nodes to combine with
other useful packets in order for the receivers to decode.
Figure 3 shows the energy consumption per packet as a
function of mobility. Among all protocols, Classical
Flooding consumes the most energy due to its flooding
nature, while CodeBCast consumes the least energy. SMF
and EBOLSR perform in between. Figure 4 shows the
network lifetime as the function of mobility speed. Among
all the protocols, Classical Flooding has the shortest
lifetime since more nodes in the network rebroadcast,
regardless of their residual energy levels. The network
lifetimes of SMF, CodeBCast and EBOLSR are similar.
From Figure 3, we notice that the energy consumption per
packet of EBOLSR and SMF is higher than CodeBCast,
while the network lifetime of the three protocols is similar.
This is because the energy consumption per packet is
defined as the total consumed energy divided by the
number of packets successfully delivered. And CodeBCast
has higher PDR than EBOLSR and SMF. In fact, the total
energy consumption and the energy consumption of each
node of these three protocols is almost the same. As a
result, the network lifetime of them is also similar.

Figure 3. Energy consumption per packet

203
 Figure 4. Network lifetime

Figure 5 shows the average number of EMPRs per node much lower than that of the WiFi scenario. The reason is
in EBOLSR and the average number of MPRs per node in that, different from the default transmission range, 250
SMF within the whole simulation time of 600s. In both meters, in the WiFi scenario, the transmission range in this
protocols, each node selects its (E)MPR set when sending a tactical scenario is variable and also much longer. All other
Hello Packet. Since the Hello interval is the same in both nodes are 1-hop neighbors of the selector most of the time,
protocols, the total number of times a node selects which result in 0 (E)MPR.
(E)MPRs are the same as well. From the simulation results,
we observe that in the EBOLSR protocol, each node selects TABLE III. PERFORMANCE COMPARISON IN RICEANRESCUE
more relaying nodes from its 1-hop neighbors to SCENARIO
rebroadcast than that of SMF. Since all the EMPR nodes
Energy
selected will participate in the rebroadcasting, this result Average
per
Lowest
# of
can explain why EBOLSR does not outperform SMF in PDR E2E energy
packet MPRs
terms of energy consumption per packet and the network delay (s) (J)
(J)
lifetime. Although it considers the residual energy of the Flood-
candidates in selecting the relay nodes, ENOLSR ends up ing
99.99% 0.1057 0.0761 451.72
electing more nodes to rebroadcast, which results in a CodeB
higher overall energy consumption. 97.62% 0.5783 0.0198 854.21
Cast

RiceanRescue Scenario: In the RiceanRescue SMF 97.23% 0.0563 0.0206 855.85 13.86
scenario, the performance of each protocol is summarized EBOL
95.90% 0.0535 0.0226 844.22 14.38
in Table III. Since each node has enough initial energy in SR
this scenario, none will die before the simulation ends. We
TABLE IV. CONFIDENCE INTERVAL AND STATISTICALLY
therefore focus on the node with the lowest remaining SIGNIFICANT DIFFERENCES (AT 95% CONFIDENCE LEVEL)
energy in the network instead of evaluating the network
lifetime. Again, we observe that Classical Flooding Differences
provides the highest PDR, but has the highest energy CodeBCast SMF statistically
significant?
consumption per packet and the lowest remaining energy (96.81%, (95.67%,
among the four protocols. CodeBCast outperforms the PDR No
98.40%) 98.79%)
others in terms of the energy consumption per packet, but Energy per (0.0194, (0.0203,
Yes
with the longest end-to-end delay. Compared with Classical packet 0.0201) 0.0209)
Flooding and CodeBCast, SMF and EBOLSR have Lowest (852.9478, (854.5336,
No
relatively low PDR, but the shortest end-to-end delay. energy 855.4898) 857.1684)
Their energy consumption per packet is similar to that of Furthermore, by comparison, the PDR, energy
CodeBCast, but a little higher. The lowest remaining consumption per packet and the lowest remaining energy
energy of EBOLSR is lower than SMF and CodeBCast, of CodeBCast and SMF are quiet close to each other. The
and SMF outperforms all other protocols, achieving the confidence intervals and statistically significant differences
best value for the lowest remaining energy. We can observe of these metrics at 95% confidence level for SMF and
that values of the average number of (E)MPRs per node are

204
 CodeBCast are given in Table IV, based on the statistics
collected from our 10 simulation runs. The confidence
intervals for PDR and lowest remaining energy overlap,
and the sample average falls within the other protocol’s
Figure 5. Average number of (E)MPRs per node
V. CONCLUSIONS
The goal of this paper is to explore energy-efficient
protocols in broadcasting scenarios and compare a new
proposed protocol with three other broadcast routing
protocols in MANETs. We adopted the multipoint relay
selection strategy based on residual energy in the EOLSR
protocol and use it in broadcasting scenarios. The proposed
EBOLSR protocol is implemented as the combination of the
EMPR selection strategy and the SMF broadcasting method.
Through the performance evaluations and comparisons, we
observe that in both scenarios, Classical Flooding has the
highest PDR but also the highest energy consumption, while
CodeBCast has the lowest energy consumption but also the
longest end-to-end delay. EBOLSR and SMF achieve similar
results in terms of PDR, average end-to-end delay, energy
consumption and lifetime, but EBOLSR does not outperform
SMF with respect to energy efficiency, contrary to what we
expected. To explain these results, there are several valid
reasons as follows:
• Packet Size: In order to record and update the residual
energy of neighbors, more fields are added in the Hello
message to record the residual energy of node’s 1-hop
neighbors. Each field recording the residual energy is 4
bytes and the number of fields is the number of 1-hop
neighbors of the sending nodes. Since the packet size
is bigger than that in SMF, it results in more energy
consumption in the transmission and reception of
Hello messages.
• Number of Relaying Nodes: Although EBOLSR takes
neighbors’ residual energy into account when selecting
205
confidence interval, indicating that the performance
differences are not statistically significant. However,
CodeBCast’s lower energy consumed per delivered packet
is statistically significant.
the relying nodes, the number of EMPRs selected is
higher than the number of MPRs in SMF. More
rebroadcasting nodes results in more energy
consumption.
• Broadcasting Network: EOLSR achieves longer
network lifetime than OLSR, while our protocol does
not outperform SMF in this respect. The reason is that
EOLSR is a unicast routing protocol but EBOLSR is a
broadcast routing protocol. After each node selects it
EMPR set, EOLSR only selects one of relaying nodes
to forward the data packet, following its routing
strategy to deliver the data packet from the source to
the destination. In EBOLSR, however, all selected
EMPRs must rebroadcast the data packets and thus
cause more energy consumption.
Through the performance comparison of the four
protocols, we conclude that CodeBCast is a good choice to
broadcast data packets in networks which require high
throughput and low energy consumption, and SMF is
suitable for networks that require short end-to-end delay and
low energy consumption. Future work in the design of
energy-efficient broadcast protocols in MANETs should aim
to reduce the transmission redundancy and overall network
overhead, and thus achieve the minimum energy
consumption and maximum network lifetime.
ACKNOWLEDGEMENTS
This work has been supported partially by an NSERC
Discovery Grant (#RGPIN42878) and the Communications
Research Center under the Intergovernmental Agreement
#8003930.
REFERENCES
[1] D. Johnson, Y. Hu, D. Maltz, Dynamic Source Routing Protocol for
Mobile Ad Hoc Networks for IPv4, RFC 4728
[2] C. Perkins, E. Belding-Royer, S. Das, Ad hoc On-demand Distance
Vector (AODV) Routing, RFC 3561
[3] S. Mahfoudh, P. Minet, EOLSR: An Energy Efficient Routing Based
on OLSR in Wireless Ad Hoc and Sensor Networks, Proc. of the
22nd International Conference on Advanced Information Networking
and Applications, Mar. 2008, pp. 1253-1259.
[4] J. Macker, Simplified Multicast Forwarding, draft-ietf-
manet-smf-09.
[5] L. Li, R. Ramjee, M. Buddhikot, S. Miller, Network Coding-
Based Broadcast in Mobile Ad hoc Networks, 26th IEEE
International Conference on Computer Communications,
IEEE, May 2007, pp. 1739-1747.
[6] Y. C. Tseng, S. Y. Ni, Y. S. Chen, and J. P. Sheu, The broadcast
storm problem in a mobile ad hoc network, Wireless Networks,
Volume 8, Issue 2-3, 2002, pp. 153-167.
[7] S. Guo, O. Yang, Energy-aware multicasting in wireless ad hoc
networks: A survey and discussion, Computer Communications,
Volume 30, Issue 9, Jul. 2007, pp. 2129-2148.
[8] J. E. Wieselthier, G.. D. Nguyen, A. Ephremides, On the construction
of energy efficient broadcast and multicast trees in wireless networks,
Proc. of IEEE INFOCOM, 2000, pp. 585-594.
[9] A. K. Das, R. J. Marks, M. A. El-Sharkawi, P. Arabshahi, A. Gray,
MDLT: a polynomial time optimal algorithm for maximization of
time-to-first-failure in energy-constrained broadcast wireless
networks, IEEE Global Telecommunications Conference, Dec. 2003,
pp. 362-366.
[10] I. Kang, R. Poovendran, Maximizing static network lifetime of
wireless broadcast ad hoc networks, IEEE International Conference
on Communications, Volume 3, May 2003, pp. 2256-2261.
206
[11] S. Guo, O. Yang, Maximizing multicast communication lifetime in
wireless mobile ad hoc networks, IEEE Transaction on Vehicular
Technology, Volume 57, Number 4, July 2008, pp. 2414-2425.
[12] O. Egecioglu, T.F. Gonzalez, Minimum-Energy Broadcast in Simple
Graphs with Limited Node Power, Proc. of International Conference
on Parallel and Distributed Computing and Systems, Anaheim, CA,
Aug. 2001, pp. 334-338.
[13] M. Cagalj, J.P. Hubaux, and C. Enz, Minimum-Energy Broadcast in
All-Wireless Networks: NP-Completeness and Distribution Issues,
Proc. of 8th Annual International Conference on Mobile Computing
and Networking, Atlanta, Georgia, Sep. 2002.
[14] W. Liang, Constructing Minimum-Energy Broadcast Trees in
Wireless Ad Hoc Networks, Proc. of 3th ACM International
Symposium on Mobile Ad Hoc Networking and Computing, Lausanne,
Switzerland, Jun. 2002, pp. 112-122.
[15] J.E. Wieselthier, G.D. Nguyen, and A. Ephremides, On the
Construction of Energy-Efficient Broadcast and Multicast Trees in
Wireless Networks, Proc. of the IEEE Conference on Computer
Communications, New York, Jun. 2002.
[16] I. Kang, R. Poovendran, A Novel Power-Efficient Broadcast Routing
Algorithm Exploiting Broadcast Efficiency, Proc. of IEEE Vehicular
Technology Conference, Orlando, Oct. 2003, pp. 2926-2930.
[17] P. Jacquet, T. Clausen, Optimized Link State Routing Protocol
(OLSR), RFC 3626
[18] NRL SMF project, http://cs.itd.nrl.navy.mil/work/smf/index.php
[19] T. Kunz, Efficiently Supporting One-to-Many and Many-to-Many
Communication Patterns in Narrowband Tactical Networks:
Flooding, Efficient Broadcasting, and Network Coding, Technical
Report, Mar. 2009.
[20] S. Guo and O. Yang, Localized Operations for Distributed Minimum
Energy Multicast Algorithm in Mobile Ad Hoc Networks, IEEE
Trans. Parallel and Distributed Systems, Feb. 2007, pp. 186-198.