QoS Topology Control with Energy Efficiency for MANET

Fujian Qin
Journal of Convergence Information Technology, Volume6, Number 6, June 2011

QoS Topology Control with Energy Efficiency for MANET

Fujian Qin
College of Computer Science, Chongqing University of Arts and Sciences, Chongqing, China,
qinfujian@163.com
doi : 10.4156/jcit.vol6.issue6.31

Abstract
Nowadays, a number of modern applications require Ad Hoc network to provide QoS guarantee
and it has received much attention. In this paper, we propose an efficiently QoS topology control
algorithm. Our algorithm dynamically adjusts transmission power of mobile nodes to construct new
topology which can meet bandwidth and end-to-end delay constraints as well as minimize the total
energy consumption in network. It can be modeled as a problem of multi-commodity flows and easily
be implemented. Simulation results show that the algorithm has a better performance and will provide
an effective solution for realizing multimedia applications in mobile network.

Keywords: MANET, QoS Topology Control, Energy Efficiency

1. Introduction
The mobile ad hoc network (MANET) is a collection of wireless mobile nodes that form a temporary
network without the use of a fixed infrastructure or centralized administration. Each node acts as a host
as well as a router. MANETs are found in applications such as battlefields, major disaster areas, and
short-term activities [1]. With most multimedia and other time or error sensitive services increasing,
Quality of Service (QoS) provisioning and energy saving become more and more important [2]. The
provision of QoS in MANET, however, is much more challenging than in wire-line network, mainly
due to node mobility, multi-hop communications, contention for channel access, and a lack of central
coordination [3]. The difficulties in the provision of such guarantees have limited the usefulness of
MANETs. The QoS routing protocol is an integral part of any QoS solution since its function is to
ascertain which nodes, if any, are able to serve requirements for applications. The goals of QoS routing
are selecting routes with satisfied QoS requirement and achieving global efficiency in resource
utilization [4]. Consequently, it also plays a crucial role in data session admission control. Energy
saving for mobile nodes is another critical issue since if a nodes battery is drained, it cannot function
at all. Node failures can also cause network partitioning, leading to a complete network failure and no
service provisioning at all. Hence, power aware and energy efficient MAC and routing protocols have
received a great deal of research attention.
Note that the end-to-end delay and bandwidth are often dependent on the network topology.
Inappropriate topology will cause some mobile nodes become the bottleneck of resources and reduce
the probability for discovering QoS routing. Over the last several years, many researchers begin to
consider adjusting the transmission power of nodes to construct a topology which can meet to QoS
requirements and the total transmission power of nodes is minimized. Wieselthier and Nguyen [5]
introduce and evaluate algorithms for tree construction in infrastructureless, all-wireless applications.
Energy efficiency is used as the performance metric to evaluate broadcast and multicast trees. They
develop the broadcast incremental power algorithm and adapt it to multicast operation, which exploits
the broadcast nature of the wireless communication environment, and addresses the need for energyefficient operation. Cheng et al. [6] consider the approximated solutions for the minimum energy
network connectivity problem in MANET. They present a theorem that reveals the relation between the
energy consumption of an optimal solution and that of a spanning tree, and propose an optimization
algorithm to improve the result of any spanning tree-based topology. Furthermore, two polynomial
time approximation heuristics are provided to compute the power assignment of wireless nodes in both
static and low mobility wireless networks. Cardei et al. [7] address the topology control with
cooperative communication (TCC) problem in MANET, which aims to obtain a strongly-connected
topology with minimum total energy consumption. Since the TCC problem is NP-complete, they

- 300 -

QoS Topology Control with Energy Efficiency for MANET

Fujian Qin
Journal of Convergence Information Technology, Volume6, Number 6, June 2011

design two distributed and localized algorithms to be used by the nodes to set up their communication
ranges. Both algorithms can be applied on top of any symmetric, strongly-connected topology to
reduce total power consumption. Hai et al. [8] consider two cases of the QoS topology control problem:
the traffic demands are not splittable as well as the traffic demands are splittable, which are formulated
as the integer linear programming problem and the mixed integer linear programming problem
respectively. They propose a greedy algorithm and an approximation algorithm with ratio n are
proposed to solve the problem, where n is the number of nodes in MANET. Cabrera et al. [9] present
the first control theoretic investigation of topology control in MANET. They first take a simple
representative fully distributed local information no topology algorithm and show that it is unstable
under certain conditions. After that, they formulate such algoirthm in a control theoretic context, and
derive a new mechanism, which is shown to be stable for a wide range of parameter variations.
In the process of topology controlling, if we do not optimize the transmission power, some nodes
probably consume excessive energy, which can result in the reducing the life-span of network. To
address the problem, I we propose new QoS topology control algorithms that will meet the given delay
and bandwidth constraints and at the same time, minimize the total transmission power for mobile
nodes.

2. System model
In this section, we introduce the system model by layered protocol stack and define our research
problem.
At physical layer, we adopt cooperative communication that combines partial signals containing the
same information to obtain complete information. Single-antenna nodes in a multi-user scenario can be
used to share their antennas to create a virtual multiple-input multiple-output system, which can
combine advantages, including the power savings provided by multi-hopping, the spatial diversity
provided by the antennas of separate mobile nodes and node cooperation can also lead to increased data
rates [10]. Since no synchronization is required, the receiver can assemble multiple copies of the same
packet received at different time. Note that there are two thresholds on the received signal SNR, i.e.,
successful decoding and timing acquisition. The combining process can be extended to multiple partial
receptions. The channel gain is often modelled as a power of distance, while the header of a message is
coded in a different way that requires a lower SNR to decode. Thus the combiner knows which packet
a partial reception belongs to. Signal combining can be performed whenever new partial reception is
made in an incremental way. Many partial receptions of a packet do not require extra storage space. By
effectively using partial signals, a packet can be delivered with less transmission power and optimize
the topology control problem.
It is difficult to guarantee correct code assignment at all time due to the continuously changing
network topology. Moreover, since not every node is active at all times, it may be desirable to
oversubscribe the medium by assigning the same code to two neighboring terminals, thus violating the
assignment goal [11]. At MAC layer, when a terminal sends a RTS, it inserts in that RTS the identity
of the code that such terminal intends to use for the ensuing data packet. A neighborhood that is
receiving a packet on the same code can then respond back with the CTS, which prevents the terminal
from commencing its data transmission. It allows terminals to be informed about all neighborhood
activities. Additionally, we can combine the code assignment and access schemes, in which the
RTS/CTS handshake over the common channel serves to reserve codes so that while the reception is
ongoing, no other neighboring terminal use any of the reserved codes. The motivation is to reduce the
overhead of exchanged information sharing. That is, information distributed to solve one problem can
be used to solve the other one. Furthermore, the MAC layer represents the most dynamic and mobilitytransparent layer of the protocol stack. Thus, it is beneficial to do code assignment at the MAC layer.
On the other hand, separating the two problems has its own advantages, including fairness. Since it is
generally difficult to provide fairness in a contention-based MAC protocol, an upper layer code
assignment can account for that.
At network layer, we adopt the proactive protocol DSR because it can collect topology information
even when it is not required to perform packet forwarding. Such protocol uses the apparently
unnecessary information to facilitate other layers tasks.

- 301 -

QoS Topology Control with Energy Efficiency for MANET

Fujian Qin
Journal of Convergence Information Technology, Volume6, Number 6, June 2011

MANET can be represented by a weighted graph G (V, E), where V is the set of nodes in the network
and E is the set of links with connected nodes. In this paper, we adopt the widely used power mode in
wireless network:
Pij=(dij)

(1)

where Pij is the power required to ensure the normal communication between node i and j, dij is the
distance between the two nodes and is a parameter which is not less than 2. Assume that each node
can dynamically adjust its power level, which can not exceed the maximal power P. Let P(i) be the
transmission power of node i. 0in, we have 0P(i)P, where n is the number of nodes in network.
According the value of transmission power we can judge whether two nodes stay connected. If P(i)
(di,j) then there exists a link between node i and j, i.e., edge (i, j)E. Let Bs,d and Ds,d denote the
n

bandwidth and delay constraints of the node pair (s, d) respectively, and ptotal P (i ) . Thus, the QoS
i 1

topology control problem in this paper can be described as follow.

Given a set of nodes V with their corresponding coordinates, Bs,d and Ds,d of node pair (s, d), where s,
dV, we need calculate the transmission power P(i) for each node i (0in) and minimize Ptotal while
meeting the end-to-end QoS constraints.

3. QoS topology control algorithm for energy efficiency

Our algorithm aims to construct a topology so that all traffic flows can meet the QoS requirements
and decrease the total transmission power in network as more as possible. If the traffic on the given
node pair can be divided along different paths, it is actually a problem about multi-commodity flows.
For a given topology, the problem is to find the routing for flows and at the same time to minimize the
maximal load of nodes. If the optimal solution of the topology exists, it means that we can find the
routing meeting the QoS constraints. Otherwise, the topology will be considered unsuitable for QoS
requirement of the flows and it need to renegotiate for QoS.

3.1. QoS formulation

Given a topology graph G (V, E) and the QoS constraints, Bs,d and Ds,d, of the given node pair, we
need discover the QoS routing for the flows and minimize the maximal load of nodes, that is, the sum
of bandwidth requirement for passing them. The mathematical model of the problem can be described
as follows:

Min( Bmax )

f i,s,jd

( f
( s,d )

f js,,id

s,d
i, j

Bs ,d

Bs ,d

(2)

(s i)
(d i )
otherwise

f js,,id ) Bmax

i, j V

(3)

(4)

1
Bs , d

s,d
i, j

Ds , d

i, j V

(5)

(i , j )

f i ,s j, d 0 Bmax 0

- 302 -

i, j V

(6)

QoS Topology Control with Energy Efficiency for MANET

Fujian Qin
Journal of Convergence Information Technology, Volume6, Number 6, June 2011

Note that for any node pair (s, d), if (i, j)E then f i , j 0 , where f i , j
s ,d

s ,d

denotes the sum of

bandwidth required along link (i, j) between node s and d. formula (2) is the optical function, which
aims to minimize the bandwidth flows require. formula (3) ensure that all links on (s, d) should meet
the flows conservation. formula (4) requires that the nodes passed by most flows should meet the
bandwidth constraints. formula (5) gives the delay constraints. Because the flows can choose different
paths, the average hops multiple flows travel through can be defined as delay. This is a standard linear
programming problem and the optimal solution can be found within polynomial time.

3.2. Energy efficient topology control

Once the Bmax calculated in previous section is greater than the available bandwidth B, it means that
the bandwidth requirement will not be met for some nodes in network. Obviously, it breaks down the
bandwidth constraint and shows that the current network topology can not be available for the QoS
traffic. Thus we need continually add some new links in the network until BmaxB. Two energy
efficient QoS topology control algorithms are provided in the paper.

3.2.1. MST based QoS topology control algorithm

First we consider a minimum spanning tree (MST) for QoS topology controlling. Since we aim to
reduce the total transmission power, the minimal power will be chosen to add a new link. Firstly, all
node pairs are sorted in ascending order according to the distances between nodes. Then choose the
shortest link which does not exist in the current topology and increase transmission power of the node
until it can connect with another node along the link. Finally, calculate the formulas defined in section
3.1 to determine whether the flows can meet the QoS requirement in the new topology. If not the case,
iterate this operation until find a QoS routing or all nodes in the network reach their maximal
transmission power P. The algorithm is described step by step in the following way:
Step 1 Sort edges in E by length in nondecreasing order and initialize T=;
Step 2 For each edge (u, v)E in the sorted order, if the neighbor sets of u and v are not identical,
make u connected with v in T. Randomly choose a vertex r as the root of spanning tree T;
Step 3 Implement the QoS algorithm to obtain Bmax in T. If BmaxB, it means that the available QoS
topology is found. Record the P(u) for node u and terminate algorithm. Otherwise, goto step
4;
Step 4 If the transmission powers of all nodes have reached their maximal P, report that there is no
available QoS topology and terminate algorithm. Otherwise, call criticalPathReduction(T)
and goto step 2.
Note that the function criticalPathReduction() is used to further reduce the cost of the minimum
spanning tree T as described in [6]. The basic idea of it is to create cycles on critical paths rather than to
use the backtracking edges e(v) if doing this can reduce the cost.

3.2.2. MIP based QoS topology control algorithm

Below we consider a minimum incremental power (MIP) tree for QoS topology controlling. Similar
to MST, new nodes are added to the tree one at a time until all nodes are included in the MIP tree.
However, it is based on the standard Prim algorithm with fundamental difference. Whereas the inputs
to Prims algorithm are the link costs, MIP must dynamically update the costs at each step to reflect the
fact that the cost of adding new nodes to a transmitting nodes list of neighborhoods is the incremental
cost. Considering node u is already in the tree and node v is not yet in the tree, the pair (u,v) which
results in the minimum value of the incremental transmission power will be selected. The algorithm is
described step by step in the following way:
Step 1 Randomly choose a vertex r from V. Initialize T= and vertex set S={r};
Step 2 For uS and vV-S, calculate the power increment

- 303 -

QoS Topology Control with Energy Efficiency for MANET

Fujian Qin
Journal of Convergence Information Technology, Volume6, Number 6, June 2011

pu ,v | (u , v) | max{0,| (u, v) | p(u )} , where p(u ) max (u , x )T | (u , x) |

Choose the node pair (u,v) with the minimum power increment pu,v. Make u connected with
v in T and let S=Sv;
Step 3 Implement the QoS algorithm to obtain Bmax in T. If BmaxB, it means that the available QoS
topology is found. Record the P(u) for node u and terminate algorithm. Otherwise, goto step
4;
Step 4 If the transmission powers of all nodes have reached their maximal P, report that there is no
available QoS topology and terminate algorithm. Otherwise, call criticalPathReduction(T)
and goto step 2.
Note that step 4 is the same as that of MST based algorithm.

4. Experimental results
In this section, we conduct simulation experiments to evaluate the performance of the two
algorithms (called MSTQTC and MIPQTC respectively) proposed in section 3.2. Mobile nodes are
arbitrarily placed in a simulation area 10001000. Initially the available bandwidth B of all nodes is set
to 500. k node pairs will be randomly generated in the simulation process. The bandwidth constraint
Bs,d of (s, d) follows normal distribution N(Bm, 0.25Bm), where Bm and 0.25Bm are the mean and
variance of QoS requirement respectively. The delay constraint for all node pair can be simply set to
2n/3, where n is the number of nodes in network. The parameter in the power model is set to 2 and 4.
Empirically, Bm=50 and P=(1000 2 )2.
Fig. 1 plots the total energy consumption with the size of network changing when is 2. As is
shown in the figure, the energy consumptions of both algorithms decrease with the number of nodes
increasing. This is due to the fact that the nodes density increases correspondingly and each node can
reduce its transmission power to reach other nodes. However, the performance of MIPQTC has always
been higher than that of MSTQTC because it always chooses the minimal transmission power to create
a new link. In the best case, MIPQTC consumes 60 percent less total energy than MSTQTC.
Furthermore, we can find that when the value of n is small, the total energy consumption of two
algorithms increases fast. The reason is that the high node density provides more choices for flows and
therefore reducing the energy consumption of finding a QoS routing.

Figure 1. Energy consumption versus n (=2)

Fig. 2 plots the total energy consumption with the size of network changing when is 4. As is
shown in the figure, the advantage in power efficiency by using MIP solution still holds. However, the
difference between power consumption of the two algorithms is less distinctive. In the best case,
MIPQTC consumes 17 percent less total energy than MSTQTC.

- 304 -

QoS Topology Control with Energy Efficiency for MANET

Fujian Qin
Journal of Convergence Information Technology, Volume6, Number 6, June 2011

Figure 2. Energy consumption versus n (=4)

Fig. 3 plots the average out degree of nodes with bandwidth B m of node pair changing.
Empirically, the value of Bm varies from 20 to 200, n=30 and P=(1000 2 ) 2 . Because the
constructed network topology is directed, the average out degree of nodes is an important
indicator to evaluate the transmission quality in network. As is shown in the figure, the
performance of MSTQTC is superior than that of MIPQTC. The flows can choose more paths
for routing while B m growing, the average out degree of nodes subsequently increases in both
algorithms. Since MIPQTC always chooses the smallest transmission power of nodes while
adding a new link into topology, it probably makes the same node continually increasing its
power and have greater out degree. With the B m increasing, however, the average out degree of
MSTQTC grows faster. Thus, only if the bandwidth requirements are very large, MIPQTC
algorithm will be more favorable.

Figure 3. Average out degree versus Bm

Fig. 4 plots the average out degree of nodes with the maximum transmission range changing. Again,
it shows that the performance of MSTQTC algorithm is significantly better than that of MIPQTC
since it causes imbalance of transmitting power of nodes. The imbalance of transmitting power
of nodes would make some nodes reach their maximal power quickly. Note that the average out
degree of nodes does not always decrease with transmission range increasing. This is because some
nodes reach their maximal power level and consume too much energy when transmission range is too
large, but it also consume much energy if there are a lot of hops between node pairs when it is too
small.

- 305 -

QoS Topology Control with Energy Efficiency for MANET

Fujian Qin
Journal of Convergence Information Technology, Volume6, Number 6, June 2011

Figure 4. Average out degree versus transmission range

5. Conclusions
In this paper, we have discussed the energy efficient QoS topology control problem. For each node
pair in network, once the bandwidth and delay can not be met, we can adjust the transmission power of
nodes and add new links into the topology so as to meet the QoS requirements. Two algorithms are
proposed to construct such QoS topology. Simulation results show that the algorithms can effectively
reduce the total energy consumption of the network and achieve better performance especially in the
scene of moderate bandwidth constraints.

6. Acknowledgements
We are grateful to the referees for their valuable comments and suggestions. This work is supported
by the Scientific Research Fund of Chongqing Municipal Education Commission under grant
KJ111218 and the Research Project of Chongqing University of Arts and Sciences under grant
Z2010JS18.

7. References
[1] Mamoun Hussein Mamoun, A New DSR Routing Protocol for MANET, Journal of
Convergence Information Technology, Vol. 4, No. 4, pp. 27 - 30, 2009.
[2] Yuan Li, Xinmeng Chen, Dan Yu, Bandwidth- Guaranteed Multi-Path QoS Routing for Saving
Energy in TDMA-Based Wireless Ad Hoc Networks, In proceedings of Wireless
Communications, Networking and Mobile Computing, pp. 1-4, 2006.
[3] Lajos Hanzo, Rahim Tafazolli, A Survey of QoS Routing Solutions for Mobile Ad Hoc
Networks, Communications Surveys & Tutorials, pp. 50-70, 2007.
[4] Zhang Baoxian, HT Mouftah, QoS routing for wireless ad hoc networks: problems, algorithms,
and protocols, Communications Magazine, pp. 110-117, 2005.
[5] Qi Yang, Jianghong Shi, Biyu Tang, Distributed Dynamic Channel Access Scheduling Method
for Wireless Ad Hoc Network, Journal of Digital Content Technology and its Applications, Vol.
5, No. 3, pp. 310-319, 2011.
[6] Cheng Cardei, Sun Jinhua, Topology Control of ad hoc Wireless Networks for Energy
Efficiency, IEEE Transactions on Computers, Vol. 53, No. 12, pp. 1629 - 1635, 2004.
[7] Mihaela Cardei, Jie Wu, Shuhui Yang, Topology control in ad hoc wireless networks using
cooperative communication, Mobile Computing, Vol. 5, No. 6, pp. 711 - 724, 2006.
[8] Liu Hai, Li Deying, Jia Xiaohua, QoS Topology Control with Minimal Total Energy Cost in Ad
Hoc Wireless Networks, In proceedings of The Second International Conference on Mobile Ad
Hoc and Sensor Networks, pp. 13-15, 2006.

- 306 -

QoS Topology Control with Energy Efficiency for MANET

Fujian Qin
Journal of Convergence Information Technology, Volume6, Number 6, June 2011

[9] Joao Cabrera, Ram Ramanathan, Carlos Gutierrez, Stable Topology Control for Mobile Ad-Hoc
Networks, Communications Letters, Vol. 11, No. 7, pp. 574 - 576, 2007.
[10] Jie Wu, Mihaela Cardei, Extended Dominating Set and Its Applications in Ad Hoc Networks
Using Cooperative Communication, IEEE Transactions on Parallel and Distributed Systems, Vol.
17, No. 8, pp. 851-864, 2006.
[11] Alaa Muqattash and Marwan Krunz, CDMA-Based MAC Protocol for Wireless Ad Hoc
Networks, In proceedings of MobiHoc03, pp. 153-164, 2003.

- 307 -