Globecom 2013 - Communications Software, Services and Multimedia Symposium

Video Quality-Maximizing Resource Allocation and

Scheduling with Statistical Delay Guarantees
Amin Abdel Khalek, Constantine Caramanis, and Robert W. Heath Jr.
Wireless Networking and Communications Group
The University of Texas at Austin
2501 Speedway Stop C0806, Austin, Texas 78712-1687
Email: akhalek@utexas.edu, caramanis@mail.utexas.edu, rheath@ece.utexas.edu

AbstractReal-time video demands quality-of-service (QoS)

guarantees such as delay bounds for end-user satisfaction. Due
to the stochastic nature of wireless fading channels, deterministic
delay bounds are prohibitively difficult to guarantee. Instead, this
paper proposes providing statistical delay guarantees using the
concept of effective capacity. A multiuser setup is considered
whereby different users have (possibly different) delay QoS
constraints. The resource allocation policy that maximizes the
sum video quality is derived and has a useful intuitive interpretation: The optimal operating point per user is such that
the rate-distortion slope is the inverse of the supported video
source rate per unit bandwidth, termed source spectral efficiency.
Scheduling policies are also proposed to select a maximal user
subset such that all selected users can meet their statistical
delay requirement. Results show that QoS-aware scheduling and
resource allocation enable supporting significantly more users
under the same resource constraints.

I. I NTRODUCTION
The real-time nature of video transmission requires maintaining stringent delay bounds to ensure a good user experience. The stringency of the delay bound is further dependent
on the specific use case. For instance, interactive applications
such as video conferencing can only tolerate an end-to-end
delay in the order of few hundred milliseconds for a smooth
experience whereas with live streaming, the delay constraint
can be relaxed to few seconds. Since the bottleneck in the endto-end network delay is queuing, users with more stringent
delay constraints should be allocated more physical resources
to boost their service rates, reduce the queuing delay, and
support their QoS requirement. Furthermore, there is an inherent tradeoff between queuing delay and video source rate.
Therefore, rate distortion optimization should be applied to
ensure that satisfying delay constraints does not come at the
expense of maintaining high perceptual quality.
In this paper, we derive a resource allocation and rate
adaptation policy that maximizes the sum video quality such
that all users in the network can achieve a target statistical
delay bound. Statistical delay bounds are considered since
deterministic delay bounds are prohibitively expensive to
guarantee over wireless networks in the presence of fading.
We show that the optimal operating point per user is such that
the rate-distortion slope is the inverse of the supported video
source rate per unit bandwidth. The maximum source rate per
unit bandwidth is a fundamental measure of the number of
video bits per channel use that can be delivered subject to
the QoS requirement and is termed source spectral efficiency.

Downloaded
from http://www.elearnica.ir
978-1-4799-1353-4/13/$31.00
2013 IEEE

For the general case where the set of users in the network
cannot all be served, we solve the problem of selecting a
maximal subset of users to schedule such that each user can
meet his target QoS requirement. We show that the optimal
scheduling policy can be obtained in polynomial time in the
number of users and it involves computing the minimum
resource allocation required by each user to support their QoS
requirement, using it as a sorting criteria, and scheduling the
first sorted users such that the sum of their minimum resource
requirement does not exceed the total available resources.
Previous work on multiuser video transmission focuses on
stored video use cases in which content can be buffered ahead
opportunistically and large delays can be tolerated. For example, for stored video streaming, in [1], the knowledge of the
variations in the dependence of perceived video quality to the
compression rate is utilized for resource allocation to improve
video QoE across multiple users. In [2], a scheduling scheme
for streaming pre-encoded video over wireless networks is
presented in which rates are dynamically adjusted based on
channel quality. For real-time video, however, such as live
streaming or video conferencing, whereby the content in not
pre-encoded, large buffers are not feasible and stringent delay
constraints need to be guaranteed. For that important use case,
no previous work addresses the problem of multiuser resource
allocation under statistical delay constraints.
The effective capacity link layer model [3] characterizes the
capacity of the wireless channels in the presence of queues
using QoS exponents that describe the decay rate of the
queue length tail probability and characterize a corresponding
statistical delay bound. The notion of statistical QoS is tied
back to the well developed theory of effective bandwidth
[4][6] and its dual concept of effective capacity [3], [7],
[8]. The main limitation of previous work is that effective
capacity has not been applied previously in the context of a
cross-layer design to optimize a video quality-based utility
function. In [9], effective capacity analysis is applied for
scalable video transmission to minimize resource utilization
over unicast and multicast links. Considering energy-efficiency
as a target objective in an ad-hoc network, [10] derives energyefficient transmission schemes such that the end-to-end delay
bounds are satisfied. Our proposed approach in this paper,
on the other hand, provides a comprehensive framework to
jointly capture the tradeoff between end-to-end delay, video
perceptual quality, and resource allocation.

1736

Globecom 2013 - Communications Software, Services and Multimedia Symposium

2

Rate-Distortion
Characteristics

Delay QoS Requirements

Q1(R1)

Q2(R2)

User 1

B1

User 2

B2

R1

....

QK(RK)

....

R2

User K
RK

BK

Rate Adaptation and Resource Allocation

Rate-distortion characteristics
Delay QoS requirements
Channel state information

Fig. 1. Proposed system block diagram.

II. S YSTEM M ODEL

Consider the downlink of a base station where K video
users are orthogonally sharing a bandwidth BPHz. Each user
K
k is allocated a bandwidth Bk Hz such that k=1 Bk = B.
The individual user bandwidths are assumed narrowband so
that the corresponding channel experiences flat fading. The
channel coherence time is T . Further, the timescale of video
rate adaptation is much larger than T . This is typically the
case because the rate of the video source is adapted at the
GoP timescale which is in the order of a second. On the other
hand, the channel variation is in the milliseconds timescale.
Since resource allocation and rate adaptation are done at the
source variation timescale and the video source experiences
the ergodic capacity of the channel, we only require channel
distribution information (CDI). Let fk () represent the SNR
distribution for user k. Moreover, {k } is modeled as an ergodic and stationary block-fading process uncorrelated among
consecutive packets j. Thus, the instantaneous transmission
rate for user k is Ck = Bk log(1 + k ).
The video segment intended for user k is transmitted at
rate Rk . A queue is inserted for each user stream to absorb the mismatch between the arrival and service rate due
to the channel variations. The video stream for each user
k is characterized by a rate distortion model Qk (Rk ) that
determines the mapping from the source rate to the video
perceptual quality. The rate-quality mapping function Qk (.)
is concave and continuous. In practice, rate distortion curves
corresponding to practical video codecs always follow this
concave behavior because diminishing returns in quality are
achieved as the rate increases.
III. S TATISTICAL D ELAY B OUNDS
In this section, we describe the procedure for providing
delay guarantees by characterizing link-level QoS metrics
according to the effective capacity link layer model.
A. Queuing Model for Video Transmission
A separate queue is maintained for each video stream at the
base station. Given the SNR distribution fk () for each user,
we are interested in adaptively configuring the source rates Rk
and the bandwidth allocation Bk such that the following QoS
constraint is satisfied


Pr Dk > Dkth Pkth k

(1)

where Dk is the queuing and transmission delay for user k

video stream, Dkth is the statistical delay-bound, and Pk,th is
the target delay-bound violation probability.
For an ergodic and stationary arrival and service processes,
the queue length at time t of each queue can be bounded
exponentially as t [11]

.
th
Pr qk (t) > qkth = ek qk ;
0kK
(2)
where qk is the queuing delay at queue k and qkth is the queuelength threshold. The parameter k , termed the QoS exponent,
determines the decay rate and is used to characterize delay.
More stringent QoS requirements are characterized by larger
k while looser QoS requirements require smaller k .
B. Effective Capacity for Statistical Delay Bounds
Physical-layer channel models present a limitation for QoS
support due to the difficulty of using them in analyzing queues.
In contrast, the effective capacity channel model [3], [9]
captures a generalized link-level capacity notion of the fading
channel by characterizing wireless channels in terms of linklevel QoS metrics, such as delay-bound violation probability.
We denote by Ck (k ) the effective capacity for the k th user.
Given an service process {Ck }, its effective capacity, denoted
by Ck (k ), is defined as the maximum constant arrival rate
which can be supported by {Ck } subject to the specified
QoS exponent k . The effective capacity can be thought of
as the capacity of the underlying channel from the perspective
of upper protocol layers. The effective capacity theory states
that considering the delay as the performance metric instead
of the queue length in (2), we can write the QoS guarantee
equivalently as a function of the effective capacity

.
th
Pr Dk (t) > Dkth = ek Ck (k )Dk ;
1 k K (3)
Moreover, for a stationary and ergodic service process {Ck }
that is uncorrelated across time frames, the effective capacity
can be expressed as follows [12]


1
Ck (k ) = ln E {ek Ck }
(4)
k
where Ck = Bk log(1 + k ) is the service rate. For the QoS
constraint (1) to be met, we need to ensure that Ck (k )
Ck = Rk T . Thus, the source rate is upper-bounded as follows


1
Rk
ln E {ek Bk T log(1+k ) }
(5)
T k

1737

Globecom 2013 - Communications Software, Services and Multimedia Symposium

3

IV. Q UALITY-M AXIMIZING R ESOURCE A LLOCATION

In this section, we formulate and solve the problem of
resource allocation and rate adaptation for maximizing the
sum video quality across users subject to the statistical delay
constraint. The problem is formulated as follows

maxR,B
s.t.

K
X
k=1
K
X

Qk (Rk )

(6)

Bk = B; Bk 0 k

(7)

Theorem 1 provides the optimal resource allocation solution

by showing that the optimal operating point per user is such
that the rate-distortion slope is the inverse of the supported
video source rate per unit bandwidth. The maximum source
rate per unit bandwidth is a fundamental measure of the number of video bits per channel use that can be delivered subject
to the QoS requirement, termed source spectral efficiency.
Theorem 1. Sum Quality-Maximizing Resource Allocation:
The optimal bandwidth allocation Bk and source rate Rk for
each user k is such that
Qk (Rk (Bk , k ))

Rk (Bk , k )
|
{z
}

k=1


Pr Dk > Dkth Pkth k

(8)

where R is the source rate vector and B is the bandwidth

allocation vector. First, we show that to maximize the sum
video quality, the statistical delay bound should be met
with equality for each user. Since Qk (Rk ) is increasing,
increasing the source rate for any user improves the objective
function.
Furthermore,
the delay bound violation probability


Pr Dk > Dkth is increasing in
 the source rate. Thus, the
optimal solution must satisfy Pr Dk > Dkth = Pkth k. This
condition reduces to Ck (k ) = Rk T . It follows that the source
rate Rk (Bk , k ) should satisfy (5) with equality.
Since the QoS exponent is also a function of the source
rate which in turn is a function of the bandwidth allocation, it
cannot be computed beforehand. Thus, we compute k jointly
with Bk as part of the optimization problem. The following
Lemma re-writes the delay constraint in terms of the QoS
exponent and the bandwidth allocation using the effective
capacity expression for an uncorrelated channel.
Lemma 1. The statistical delay constraint (8) can be written
equivalently as
n
k B k T o
= pk
(9)
E (1 + k ) ln(2)
1/D th

k
where pk = Pkth
. Furthermore, the solution to the
problem requires that k = dk /Bk where the bandwidthQoS exponent product dk is a constant dependent only on
the channel distribution and the delay constraint.

Proof: See [13], proof of Lemma 2.

Applying Lemma 1, we rewrite the resource allocation
problem equivalently as follows

minB,

K
X
k=1

s.t.

K
X




Bk T
1
kln(2)
Qk
ln E {(1 + k )
}
T k

Bk = B; Bk 0 k

(10)

k=1

n
k Bk T o
E (1 + k ) ln(2)
= pk k

(11)

The problem is convex in B and . We dont include the

convexity proof for brevity. Furthermore, a feasible point
always exists if Qk (Rk ) is defined for every Rk 0, i.e.,
given any delay bound and channel conditions, there exists a
small enough source rate such that the delay bound is met.

Rate distortion slope

Rk (Bk , k )
Bk
|
{z
}

= .

(12)

Source spectral efficiency



k B k T
where Rk (Bk , k ) = T1k ln E {(1 + k ) ln(2) } and
PK
is a constant chosen such that
k=1 Bk = B. Further,
Rk (Bk , k )/Bk is independent of Bk and is only a function
of the fading distribution fk (), the delay bound Dkth , and the
delay bound violation probability Pkth .
Proof: See [13], proof of Theorem 1.
We refer to the term Rk (Bk , k )/Bk as the source spectral
efficiency as it represents the number of video bits that can be
delivered per channel use subject to the QoS constraint. The
interpretation of the above result is that the optimal resource
allocation is such that the slope of the rate distortion curve
multiplied by the source spectral efficiency is the same for all
users. Furthermore, that constant determined by is computed such that the sum bandwidth constraint is satisfied with
equality. Note that as decreases, Bk increases for all k and
vice versa, thus reducing the problem intoP
a one-dimensional
root-finding problem involving solving for k Bk ()B = 0
to find .
Note from Theorem 1 that the source spectral efficiency
jointly characterizes the channel and the QoS requirement
since it is independent of the resource allocation Bk . It is only
a function of the channel statistics, the delay bound, and the
delay bound
probability. Specifically,
Rk (Bk )/Bk =

violation
th
th
ln 1/Pk /Dk T dk . Furthermore, ln 1/Pkth /Dkth T dk is a
fundamental measure of the maximum number of video bits
per channel use that can be delivered subject to the QoS
requirement.
V. M AXIMAL U SER S UBSET S CHEDULING
In this section, we consider the problem of selecting video
users to serve given their QoS constraint. The objective is to
select the largest subset of users such that each user in the
subset can meet their delay constraint.
Theorem 2 presents the optimal scheduling policy that
selects a maximal subset of users such that all user meet their
statistical delay constraint. To summarize, the optimal scheduling policy can be obtained in polynomial time in the number
of users and it involves computing the minimum bandwidth
required by each user to support their QoS requirement, using
it as a sorting criteria, and scheduling the first sorted users
such that the sum of their minimum bandwidth requirement
does not exceed the total bandwidth.

1738

Globecom 2013 - Communications Software, Services and Multimedia Symposium

4

Theorem 2. Maximal User Subset Scheduling: Under maximum sum quality resource allocation, the scheduling policy
to maximize the number of users that can meet their QoS
requirement is as follows. Define

4.5

(13)

4
3

3.5

10

2.5

0.5
0.1

0.01

1.5

10

0.5

Practically
Infeasible
Region

0.1
1
0.0

and let l(j) be the sorting operation in increasing order so

min
that Bl(j)
is the j th sorted element j = 1, , K, then the
maximum number of users supported N is

10

Delay bound violation probability Pth

Dkth dk T min
R
=
ln(1/Pkth ) k

Bkmin

5
Source Spectral Efficiency R /B (bps/Hz)
for = 20 dB

0.5

10

0.2

N = argmaxN

N
X

min
s.t.
Bl(j)

j=1

N
X

min
B
Bl(j)

(14)

j=1

and the corresponding scheduled users are l(1), , l(N ).

Proof: See [13], proof of Theorem 2.
Algorithm 1 uses Theorem 2 to select a maximal user subset
from a set of candidate video users with delay QoS constraints.
Among the scheduled users, resources are allocated to maximize the sum video quality using Theorem 1.
Algorithm 1 QoS-aware Scheduling and Resource Allocation.
Given K users with delay QoS requirements {Dkth , Pkth }K
k=1 , fading
distributions fk (), and total bandwidth B.
Step 1. Compute QoS-related Metrics
for k = 1 K do
th
1/Dk
k.
1)
Find pk = Pkth
2)
Given fk (), find the bandwidth-QoS
exponent
product dk =


dk T
ln(2)

Bk k k by solving E (1 + k )
= pk .
3)

Compute minimum bandwidth requirement per user Bkmin =

th
dk T
Dk
Rmin
ln(1/Pkth ) k

and minimum quality qkmin = Qk (Rkmin ).

Compute maximum
source spectral efficiency Rk (Bk )/Bk =


ln 1/Pkth /Dkth T dk .
end for
Step 2. Select Maximal User Subset
min is the j th sorted
1) Sort Bkmin with operator l(j) such that Bl(j)
element j = 1, , K
P
min
2) Compute maximum number of users N = argmaxN N
j=1 Bl(j)
PN
min B
s.t. j=1 Bl(j)
3) Schedule users l(1), , l(N ).
Step 3. Allocate Resources
1) Set = initial

Qk (Rk
(Bk
,k ))
T d
k ( )
2) Solve R
= ln(p
for Rk k = l(1), , l(N )
(B , )
k)
k
k k
3) Solve k = ln(pk )/(T Rk ) k = l(1), , l(N )

4) Solve
P Bk = dk /k k = l(1), , l(N )
5) If
Bk < B, decrease
P and repeat 2,3,4, otherwise increase and
repeat 2,3,4 until | Bk B| < .
4)

VI. R ESULTS
In this section, we present results and analysis to demonstrate the performance of the proposed scheduling and resource
allocation algorithms for real-time video transmission. First,
we present analysis of the source spectral efficiency under
different QoS requirements and channel conditions. Next, we
present results for the case of multiple users to demonstrate
the gains of QoS-aware scheduling and resource allocation.

0.5

1.5
2
2.5
Delay bound Dth (sec)

3.5

Fig. 2. Contour plot of source spectral efficiency for a Rayleigh channel with

= 20 dB vs. delay bound Dth and delay bound violation probability P th .

A. Source Spectral Efficiency under Delay Constraints

Figure 2 shows a contour plot of the maximum source spectral efficiency supported by a Rayleigh channel as a function of
the delay bound and delay bound violation probability. Note,
from Theorem 1, that the maximum
source spectral efficiency


can be expressed as ln 1/Pkth /Dkth T dk and is achieved by
the proposed rate adaptation algorithm. The main observation
from the plot is that there exists a boundary beyond which
the source spectral efficiency declines very rapidly, making
practical realization of the QoS constraint impossible since it
requires either a very large bandwidth or operating at a very
low source rate. For example, to achieve Dkth = 0.5 sec and
Pkth = 103 , we require a bandwidth 1000 times the source
rate even if the average SNR is = 20 dB. This provides a
practical insight into the range of feasible QoS constraints.
B. Multiuser Scheduling and Resource Allocation
Next, we consider a multiuser setup where there is a large
number of users to be served. We apply the scheduling solution
derived in Theorem 2 to select a user subset follows by the
corresponding resource allocation policy. We consider a cell
radius of 5 Km with users uniformly distributed in the cell.
The total system bandwidth is 20 MHz.
The total number of video users is 500. Half the video
users have a delay constraint D1th = 2 sec corresponding to
live video users and the other half have a delay constraint
D2th = 0.3 sec corresponding to a typical two-way video
conferencing user. The target delay bound violation probability
is 0.1 for both sets of users. We use the Foreman, Mobile,
and Akiyo video sequences encoded with H.264/AVC. The
GoP structure is IBP BP and the GoP duration is 16
frames. The MB size is 16 16 and we use the CIF resolution
of 352 288. The minimum rate representation of the three
different videos are Rkmin = 160, 185, and 200 Kbps.
For accurate channel modeling, we derive the average SNRs
k from the distances dk using the following relation
k = 10 log10 (Pt ) KdB 10 log10 (dk ) 10 log(N0 Bk )
(15)
where KdB is the pathloss constant, is the pathloss exponent,
and N0 is the power spectral density of AWGN noise. We fix
KdB = 21.36 dB, = 3.52, and N0 = 4 1021 W/Hz.

1739

Globecom 2013 - Communications Software, Services and Multimedia Symposium

5

300

5 Km
184 users

3.4x

Scheduling: Maximal User Subset

Resource Allocation: Max Sum Quality

260

5 Km

5 Km

Proposed

280

Number of users served

240
Scheduling: Max SNR
Resource Allocation: Equal BW

220
200

160

2.1 Km
35 users
29 users

1.5 Km
14 users

Scheduling: Max SNR

Resource Allocation: Max Sum Rate

180

Baselines

140

120
100

4.1x

80
60
40
20
0
1

10

10

10

Base station transmit power (W )

5 Km

5 Km

5 Km

5 Km

Fig. 3. (a) Number of users supported by the proposed scheduling and resource allocation algorithm in comparison to the non-QoS-aware baseline vs. base
station transmit power; (b) Users covered by the proposed Maximal user subset scheduler and maximum sum quality resource allocation for Pt = 25 W; (c)
Users covered by the baseline Max SNR scheduler baseline and equal bandwidth allocation. Red markers correspond to video conferencing sessions and blue
dots correspond to live streaming users.

We consider two baselines. In both baselines, scheduling is

based on Max SNR, that is, the users with the better channels
are favored and the maximum number of the highest SNR
users that can be supported such that they can meet their
delay constraint is scheduled. In the first baseline, the available
bandwidth is divided equally among the scheduled users. In
the second baseline, the bandwidth is allocated
PK to maximize
the sum channel rate, that is Bk = Bk / j=1 j . Figure
3a shows the number of users supported using the proposed
algorithm as well as the two baselines. As we proved in
Theorem 2, the maximal user subset scheduling algorithm
with sum quality-maximizing resource allocation outperforms
any other scheduling/resource allocation combination under
the same delay constraints. Furthermore, we observe that
significant gains are achieved in terms of the total number
of users supported in the system. At high transmit powers, up
to 3.4x improvement over the baseline is achieved. In the low
transmit power regime, 4.1x is achieved. Finally, we note that
Max sum (channel) rate resource allocation baseline performs
worse than equal bandwidth allocation baseline. This is because users with good channels end up consuming a significant
fraction of the bandwidth even if their QoS requirement is less
stringent thus taking resources away from users stringent QoS
requirements with require a larger bandwidth.
To understand how these significant gains are achieved,
Figure 3b shows that the maximal user subset scheduling
algorithm supports a total of 198 users corresponding to 184
live streaming sessions and 14 video conferencing sessions.
In contrast, in Figure 3c, Max SNR scheduling supports only
64 users corresponding to 35 live streaming sessions and 29
video conferencing sessions. Thus, in essence, the gains stem
from dropping a small fraction of users that require excessively
large amount of resources to meet their QoS requirement such
that the total number of users supported is maximized.
VII. C ONCLUSION
In this paper, we used the concept of effective capacity
to provide a framework for statistical delay provisioning
for multiple users sharing a wireless network. Sum qualitymaximizing resource allocation policies were derived and
maximal user subset scheduling was proposed to maximize

the number of scheduled users that can meet their QoS

requirement. Significant gains in capacity, measured in terms
of number of users supported in the system, are achieved due
to QoS-aware scheduling and resource allocation.
VIII. ACKNOWLEDGMENT
This work was supported by the Intel-Cisco Video Aware
Wireless Networks (VAWN) Program.
R EFERENCES
[1] V. Joseph and G. de Veciana, Jointly optimizing multi-user rate adaptation for video transport over wireless systems: Mean-fairness-variability
tradeoffs, In Proc. of IEEE INFOCOM, Mar. 2012.
[2] P. Pahalawatta, R. Berry, T. Pappas, and A. Katsaggelos, Content-aware
resource allocation and packet scheduling for video transmission over
wireless networks, IEEE Journal on Selected Areas in Communications,
vol. 25, no. 4, pp. 749759, 2007.
[3] D. Wu and R. Negi, Effective capacity: a wireless link model for
support of quality of service, IEEE Transactions on Wireless Communications, vol. 2, no. 4, pp. 630 643, July 2003.
[4] A. Kumar, D. Manjunath, and J. Kuri, Communication networking: an
analytical approach. Morgan Kaufmann Publishers, 2004.
[5] C.-S. Chang, Performance guarantees in communication networks.
Springer-Verlag London, 2000.
[6] C.-S. Chang and J. Thomas, Effective bandwidth in high speed digital
networks, IEEE Journal on Selected Areas in Communications, vol. 13,
no. 6, pp. 1091 1100, August 1995.
[7] D. Wu and R. Negi, Effective capacity-based quality of service measures for wireless networks, Mobile Networks and Applications, vol. 11,
no. 1, pp. 9199, June 2006.
[8] D. Wu, Providing quality-of-service guarantees in wireless networks.
PhD thesis, Carnegie Mellon University, 2003.
[9] Q. Du and X. Zhang, Statistical QoS provisionings for wireless
unicast/multicast of layered video streams, In Proceedings of IEEE
INFOCOM, April 2009.
[10] A. Abdel Khalek and Z. Dawy, Energy-efficient cooperative video
distribution with statistical QoS provisions over wireless networks,
IEEE Transactions on Mobile Computing, vol. PP, no. 99, 2011.
[11] C.-S. Chang, Stability, queue length, and delay of deterministic and
stochastic queueing networks, IEEE Transactions on Automatic Control, vol. 39, no. 5, pp. 913 931, May 1994.
[12] J. Tang and X. Zhang, Quality-of-service driven power and rate
adaptation for multichannel communications over wireless links, IEEE
Transactions on Wireless Communications, vol. 6, no. 8, pp. 30583068,
August 2007.
[13] A. Abdel Khalek, C. Caramanis, and R. W. Heath Jr., Delay-constrained
video transmission: Quality-driven resource allocation and scheduling, Preprint, available online: https:// webspace.utexas.edu/ aa36734/
papers/ Khalek MultiuserVideo.pdf .

1740