See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/263754885

Enhanced PF Scheduling Algorithm for LTE Downlink System

Article · February 2014

CITATIONS READS

8 2,378

3 authors:

Mohamad Ibrahim Elhadad El-Sayed Mahmoud El-Rabaie

Menoufia University Faculty of Electronic Engineering, Menouf
11 PUBLICATIONS 55 CITATIONS 630 PUBLICATIONS 3,882 CITATIONS

SEE PROFILE SEE PROFILE

Mohammed Abd-Elnaby
Faculty of Electronic Engineering, Menoufia, University, Egypt
109 PUBLICATIONS 872 CITATIONS

SEE PROFILE

All content following this page was uploaded by El-Sayed Mahmoud El-Rabaie on 31 January 2015.

The user has requested enhancement of the downloaded file.

Mobile Computing Vol. 3 Iss. 1, Fe bruary 2014 www.mc-journal.org

Enhanced PF Scheduling Algorithm for LTE

Downlink System
Mohamad I. Elhadad1 , El-Sayed M. El-Rabaie2 , Mohammed Abd-Elnaby3
Department of Electronics and Electrical Communications, Faculty of Electronic Engineering, Menoufia University,
Menouf, 32952, Egypt.
[email protected]; [email protected]; [email protected].

Abstract et al. (2009)].

A ke y feature of Long Term Evolution (LTE) syste m is the Basic LTE network elements consist of a powerful
adoption of advance d Radio Resource Manageme nt (RRM) evolved Node B (eNB) station and several User
proce dures in orde r to increase the system performance.
Equipments (UEs ) in addition to a gateway. Basic
Packe t sche duling me chanisms play a fundame ntal role ,
packet scheduling is implemented by the network
be cause they are responsible for choosing how to distribute
operator in UE and eNB station for both uplink as well
radio resources among diffe re nt stations. In this paper, a
modifie d Proportional Fair (PF) sche duling algorithm is as downlink. However, there are no rigid
propose d for capacity enhanceme nt for LTE system and specifications set by 3GPP for scheduling mechanism
compare d with the PF downlink sche duler, which is leaving the details at the discretion of service provider.
characte rize d by high fairness but with lo w spectral Packet scheduling comes under RRM and its main
e fficie ncy. Simulation results show that the propose d functionality is to decide users that would transmit
algorithm e nhances the ove rall syste m capacity and also their data on the air interface. The scheduling should
provides fairness in the distribution of the resources. The integrate fairness in terms of throughput as well as the
propose d algorithm improves the average ce ll throughput service policies to which users subscribe [Hara &
by more than 10.3%, with approximate ly the same fairness Prasad (2003), Furht et al. (2009), Holma & Toskala
le ve l (2.6% re duction) as compare d with the conve ntional PF
(2009), Ergen (2009)].
sche duling algorithm.
The design of a downlink scheduling algorithm is a
Keywords
complex procedure and presents a number of design
LTE; Packet Scheduling; Fairness; RRM challenges, such as maximization of the system
capacity and spectral efficiency, fairness approach, bit
Introduction error performances and etc. [Tran & Eltawil (2012)].
The LTE standardized by Third Generation Partner- Several algorithms have been proposed in literature to
ship Project (3GPP) [3GPP (2012)] has become the provide higher spectral efficiency with fairness
most important radio access technique for providing ensured to all its users, the scheduling algorithm in
mobile broadband to the mass market. The [Schwarz et al. (2011)], for example, tries to make this
introduction of LTE will bring significant enhance- balance using Assignment Model, which effectively
ments compared to High Speed Packet Access (HSPA) maps UEs to RBs during each TTI. The proposed
in terms of spectrum efficiency, peak data rate and algorithm in [Schwarz et al. (2011)] makes
latency. Since the initial release in 2008, a slightly modification on the best-Channel Quality Indicator
modified version (Release-9) and a complete fourth (best-CQI) algorithm, which is characterized by high
generation standard named LTE-Advanced (Release- data rates at cell level but poor fairness, trying to
provide fairness in the distribution of the resources,
10) have been developed [Chadchan & Akki (2010)].
while at the same time keeping the system capacity
LTE uses Orthogonal Frequency Division Multiple
utilization as high as possible.
Access (OFDMA) in the downlink. OFDMA divides
the frequency band into a group of mutually In [Chadchan & Akki (2013)], the authors proposed a
orthogonal subcarriers, thereby improving the system method for multi-user scheduling that operates on the
capabilities by providing high data rates, supporting boundary of the achievable multiuser rate region
multi-user diversity and creating resistance to while guaranteeing a desired long term average
frequency selective fading of radio channels [Ali & fairness. Another scheduling algorithm based on the
Zeeshan (2012), Talevski & Gavrilovska (2012), Larmo utility function has been implemented in [Kela et al.

7
www.mc-journal.org Mobile Computing Vol. 3 Iss. 1, Fe bruary 2014

(2008)], in order to improve the performance of the scheme. In particular, Single Carrier Frequency
LTE system, particularly, improving the throughput Division Multiple Access (SC-FDMA) and OFDMA are
and fairness performance. The throughput fairness used in uplink and downlink directions, respectively.
between users can be effectively controlled by Different from basic OFDM, they allow multiple
dividing the packet scheduler into a time domain and accesses by assigning sets of sub-carriers to each
a frequency domain and utilizing different algorithms individual user. OFDMA can exploit sub-carriers
in both domains, as implemented in [Wengerter et al. distributed inside the entire spectrum, whereas SC-
(2005)]. FDMA can use only adjacent sub-carriers. OFDMA is
able to provide high scalability, simple equalization,
A Generalized Proportional Fair (GPF) scheduling
and high robustness against the time-frequency
approach and its application to OFDMA with
selective nature of radio channel fading. On the other
frequency scheduling has been presented in [Li et al.
hand, SC-FDMA is used in the LTE uplink to increase
(2010)], and compared to a system without frequency
the power efficiency of UEs, given that they are
scheduling, this increases the system throughput and
battery supplied [Piro et al. (2011)]. MIMO techniques
yields an improved fairness with respect to the
can be exploited both in downlink and uplink to
allocated resources and the achieved data-rate per user.
improve transmission reliability and data rate, and it is
An adaptive proportional fair scheduling algorithm
possible to use up to a maximum of four transmit and
for LTE system has been proposed in [Xie & Hui
four receive antennas [Iosif & Bănică (2013)].
(2011)] which provides a good trade-off between
capacity and fairness, by adjusting the scheduling Radio resources are allocated in a time/frequency
priority according to individual user’s channel domain. In the time domain, they are distributed
condition. every Transmission Time Interval (TTI), each one
lasting 1 ms. Furthermore, each TTI is composed of
A typical way to find a trade-off between
two time slots of 0.5 ms, corresponding to 7 OFDM
requirements on fairness and spectral efficiency is the
symbols in the default configuration with short cyclic
use of PF scheme [Kawser et al. (2012), Parruca et al.
prefix; 10 consecutive TTIs form the LTE Frame lasting
(2013), Girici et al. (2010), Sun et al. (2006)]. In this
10 ms. In the frequency domain, instead, the whole
paper, we try to improve the performance of the LTE
bandwidth divides this into 180-kHz sub-channels,
system by improving the throughput and fairness
corresponding to 12 consecutive and equally spaced
performance of the well known PF scheduling
sub-carriers [Yaacoub (2012)]. A time/frequency radio
algorithm; then the performance of the proposed
resource, spanning over one timeslot lasting 0.5 ms in
algorithm is compared to other algorithms in the
the time domain and over one sub-channel in the
literature; followed by evaluation on the performance
frequency domain, is called Resource Block (RB) and
of the proposed algorithm via simulations and the
corresponds to the smallest radio resource that can be
conclusion is made that the proposed algorithm is
assigned to a UE for data transmission [Khan et al.
very efficient in terms of both throughput and fairness.
(2012)]. Note that, given that the sub-channel
The rest of the paper is organized as follows: in dimension is fixed, the number of sub-channels varies
Section 2 a basic background on the LTE technology is according to different system bandwidth configura-
provided. Section 3 deals with the framework for LTE tions (e.g., 25 and 50 RBs for system bandwidths of 5
Downlink scheduling. We review and discuss the and 10 MHz, respectively).
traditional PF algorithm, and then our modified PF
scheduling algorithm is presented in section 4. Scheduling in LTE System
Performance evaluation and simulation results are
At the eNB, the packet scheduler distributes radio
given in section 5 followed by the conclusion in
resources among active users in order to satisfy their
section 6.
Quality of Service (QoS) needs. Scheduling decisions
are strictly related to the channel quality experienced
Overview of LTE System
by UEs. In particular, the UE periodically measures
LTE has been designed as a highly flexible radio access this channel quality using reference symbols; then it
technology in order to support several system sends the CQI feed back to the eNB, with an uplink
bandwidth configurations (from 1.4 MHz up to 20 control message [Dahlman et al. (2007)]. The
MHz). Radio spectrum access is based on the information about the quality of the time and
Orthogonal Frequency Division Multiplexing (OFDM) frequency variant channel is exploited by the link

8
Mobile Computing Vol. 3 Iss. 1, Fe bruary 2014 www.mc-journal.org

adaptation module to select, for each UE, the most conditions on each RB, due to differences in the
suitable Modulation and Coding Scheme (MSC) at the distance and shadow fading between the base station
physical level with the objective of the spectral and the mobile terminal. In this case, for a relatively
efficiency maximization. This approach is known as long time, the channel conditions experienced by one
Adaptive Modulation and Coding (AMC) and it has mobile terminal could be worse than those
been adopted by several wireless technologies, such as experienced by other mobile terminals, and the best-
Enhanced Data for GSM Evolution (EDGE) [ Halonen CQI scheduling algorithm may ‘starve’ the mobile
et al. (2003)] and Worldwide Interoperability for terminals with the bad channel condition in the way
Microwave Access (WiMAX) [Andrews et al. (2007)]. that the mobile terminal(s) with bad channel
conditions will never be scheduled (i.e. example users
Downlink control signaling is carried by three physical
at the cell edge). In a worst case scenario, (virtually all
channels [Iosif & Bănică (2011)]; and the most
the time) only one user could be scheduled [Talevski
important from a scheduling perspective is the
& Gavrilovska (2012)].
Physical Downlink Control Channel (PDCCH), which
carries assignments for downlink resources and uplink A practical scheduler should be capable of maximizing
grants, including the used MCS as shown in Fig. 1. the over all system capacity while still satisfying some
Considering that each modulation scheme degree of fairness among the users. The proposed
corresponds to a fixed physical data rate, the link scheduling algorithm tends to distribute the resources
adaptation module establishes the maximum available fairly among different users, therefore enabling
physical data rate for each UE (based on the received fairness while at the same time trying to maximize
channel quality information) to provide an optimal system capacity performances within the cell.
resource allocation among all users.
The Enhanced PF Scheduling Scheme
The idea of PF scheme is that the past average
throughput can act as a weighting factor of the
expected data rate, so that users in bad conditions will
be surely served within a certain amount of time. Its
metric can be simply expressed as [Wu & Chu (2011),
Nonchev & Valkama (2009)]:

𝑃𝐹
𝑟𝑘𝑖 (𝑡)
𝑚𝑖,𝑘 = 𝑖 , (3)
𝑅� (𝑡)
where 𝑅� (𝑡) is the average delivered throughput to the
𝑖

𝑖 − 𝑡ℎ user until time t and can be updated using an

FIG. 1 S IMPLIFIED MODEL OF LTE PAKET S CHEDULER
exponentially weighted low-pass filter [Wang et al.
Resource allocation for each UE is usually based on (2003), Nonchev & Valkama (2009)]:
the comparison of per-RB metrics: the k − th RBs 1 1
𝑅� 𝑖 (𝑡 + 1) = �1 − � 𝑅� 𝑖 (𝑡) + 𝑟𝑖 (𝑡) . (4)
allocated to the 𝑗 − 𝑡ℎ user if its metric 𝑚𝑗,𝑘 is the 𝑡𝑐 𝑡𝑐

biggest one, i.e., if it satisfies the equation [Capozzi et Here 𝑡𝑐 is the averaging window length over which
al. (2012)]: the average delivered throughput is calculated and
𝑚𝑗,𝑘 = max �𝑚𝑖,𝑘 � (1) 𝑟𝑖 (𝑡) denotes the actually realized throughput to the
𝑖
𝑖 − 𝑡ℎ user at the previous TTI.
For example, the best-CQI scheduling algorithm aims
at maximizing the overall throughput by assigning According to [Sun et al. (2006)], for conventional PF
each RB to the user that can achieve the maximum scheduling algorithm, the 𝑘 − 𝑡ℎ RB should be
throughput in the current TTI. Its metric can be simply allocated to the 𝑗 − 𝑡ℎ UE such that:
expressed as [Capozzi et al. (2012)]: 𝑟𝑘𝑖 (𝑡)
𝑗 = arg max � �. (5)
𝑏𝑒𝑠𝑡 −𝐶𝑄𝐼
𝑚𝑖,𝑘 = 𝑟𝑘𝑖 (𝑡) (2) 𝑖 (𝑡𝑐 − 1) 𝑅� 𝑖 (𝑡) + 𝑟𝑖 (𝑡)
where 𝑟𝑘𝑖 (𝑡)
is the instantaneous throughput for the The proposed enhanced scheduling algorithm aims to
𝑖 − 𝑡ℎ user at time instant (TTI) t on the 𝑘 − 𝑡ℎ RB achieve a significant increase in the total throughput
(calculated based on the CQI). In terms of fairness, this with a slight reduction in the fairness performance
scheduling algorithm is not fair. In a practical situation, compared to the conventional PF scheduling
each mobile terminal will experience different channel algorithm.

9
www.mc-journal.org Mobile Computing Vol. 3 Iss. 1, Fe bruary 2014

The new metric of the proposed enhanced scheduling Fig. 2 and Fig. 3 present the throughput and fairness
algorithm can be written as: performance respectively of the proposed enhanced
𝑚𝐸𝑛ℎ 𝑖( ) ( ) � 𝑖( ) () scheduling algorithm compared with that of the PF
𝑖,𝑘 = log 2 �𝑟𝑘 𝑡 � − 𝛼log 2 � 𝑡𝑐 − 1 𝑅 𝑡 + 𝑟𝑖 𝑡 � . (6)
and the best-CQI scheduling algorithms.
So, according to the new metric the 𝑘 − 𝑡ℎ RB should
be allocated to the 𝑗 − 𝑡ℎ UE such that:
𝑗 = argmax log2 �𝑟𝑘𝑖(𝑡)� − 𝛼log2 �(𝑡𝑐 − 1)𝑅�𝑖(𝑡) + 𝑟𝑖(𝑡)�. (7)
𝑖

The new parameter 𝛼 introduced in the proposed

metric equation is responsible for controlling the
trade-off between throughput and fairness achieved
by the proposed enhanced scheduling algorithm. The
operating range of 𝛼 is between 0 and 1 (0<𝛼<1), when
𝛼 =1 the proposed enhanced scheduling algorithm
gives the same performance of the conventional PF FIG. 2 AVERAGE CELL THROUGHPUT FOR DIFFERENT
scheduling algorithm which is characterized by high S CHEDULING ALGORITHMS
fairness but with low spectral efficiency. On the other
hand, when 𝛼=0, the proposed enhanced scheduling
algorithm gives the same performance of the best-CQI
scheduling algorithm which is characterized by high
throughput but with poor fairness performance.

As it can be seen later in the following simulation

results section that the proper selection of the value of
parameter 𝛼 gives significant increase in the achieved
throughput with slight reduction in fairness
performance. FIG. 3 FAIRNES S INDEX FOR DIFFERENT S CHEDULING
ALGORITHMS

Performance Evaluation and Simulation It is shown that the best-CQI is obviously able to
Results maximize cell throughput, however, , it performs
unfair resource sharing since users with poor channel
The performance evaluation of the scheduling
conditions (e.g., cell-edge users) will only get a low
algorithms is done by using the LTE Link Level
percentage of the available resources (or in extreme
Simulator (LLS) [Ikuno et al. (2010)]. The main
case they may suffer of starvation). On the other hand,
simulation parameters used in this simulator are
the proposed enhanced scheduling algorithm and the
summarized in Table 1.
PF scheduling algorithm are able to guarantee high
TABLE 1 THE SYSTEM SIMULATION PARAMETERS [Ikuno et al. (2010)]. fairness level.
P arameter Value
For the proposed enhanced scheduling algorithm as
Frequenc y Band 2.14 GHz
the value of 𝛼 decreases (from 1 to 0), the average cell
S ystem Bandwidth 20 MHz
No. of User Equipments (UE) per eNB 20 UEs throughput significantly increases while the fairness
No. of eNB 9 eNBs level slightly decreases. For example, with 𝛼= 0.8, the
S imulation Length 100 TTIs proposed algorithm improves the average cell
Antenna Configuration 1 Transmit, 1 Rec eive (1X1) throughput by more than 10.3%, with approximately
UE S peed 5 km/hr the same fairness level (2.6 % reduction) as compared
Uplink Delay 3 TTIs
with the conventional PF scheduling algorithm. Also,
The Distanc e between eNBs 500 m
with 𝛼 = 0.6, the proposed enhanced scheduling
eNB's Transmit Power 40 watts
The Thermal Noise Density -174 dBm/Hz algorithm improves the average cell throughput by
Rec eiver Noise Figure 9 dB more than 18%, with 10% reduction of the fairness
level as compared with the PF scheduling algorithm.
Simulation results are presented to analyze and
compare the performance of the proposed enhanced The Average Cell Throughput and Fairness Index for
(referred to as Enh) with that of the best-CQI and the Different Scheduling Algorithms are summarized in
PF scheduling algorithms. Table 2.

10
Mobile Computing Vol. 3 Iss. 1, Fe bruary 2014 www.mc-journal.org

TABLE 2 THE AVERAGE CELL THROUGHPUT AND FAIRNESS INDEX FOR

Dahlman, Erik, Parkvall, Ste fan, Sköld, Johan, and Beming,
DIFFERENT SCHEDULING ALGORITHMS .
Pe r, 3G Evolution HSPA and LTE for Mobile Broad band,
Scheduling Algorithm Throughput (Mb/s) Fairness
PF 44.06 0.776179 Else vie r Ltd, 2007.
Enh with α=1 44.17 0.776311 Ergen, Mustafa, Mobile Broadband Including WiMAX and
Enh with α=0.8 48.61 0.755492
Enh with α=0.6 52.02 0.659927 LTE, Springe r Scie nce + Busine ss Me dia, LLC, 2009.
Enh with α=0.4 58.52 0.548118 Furht, Borko, and Ahson, Sye d A., Long Te rm Evolution:
Enh with α=0.2 69.1 0.372397
3GPP LTE radio and ce llular technology, Taylor &
Enh with α=0.0 81.54 0.136801
best-CQI 81.37 0.138275 Francis Group, LLC, 2009.
Girici, Tolga, Zhu, Che nxi, Agre , Jonathan R., and
Conclusions
Ephre mides, Anthony, “Proportional Fair Sche duling
This paper elaborates the downlink packet scheduling Algorithm in OFDMA-Base d Wire less Systems with QoS
framework in LTE by proposing a Capacity Constraints,” International Journal of Information and
Enhancement based PF Scheduling algorithm, and its
Ele ctronics Engineering, Journal of Communications and
performance is compared with the best-CQI and PF
algorithms. The modified PF algorithm allows fair Ne tworks, Vol. 12, No. 1, pp. 30 – 42, Fe b. 2010.
distribution of available resources with the increasing Halone n, Timo, Rome ro, Javier and Me le ro, Juan, "GSM,
spectral efficiency. Simulation results presented in the GPRS, and Edge Pe rformance : Evolution Towards 3G /
paper show that the implementation of this modified UMTS", John Wile y & Sons Ltd, 2003.
PF algorithm enables improvement of the overall
Hara, Shinsuke , and Prasad, Ramjee, Multicarrie r Techniques
system capacity and also provides fairness in the
distribution of the resources. for 4G Mobile Communications, Boston London: Artech
House Unive rsal Pe rsonal Communications Serie s, 2003.
REFERENCES Holma, Harri, and Toskala, Antti, LTE for UMTS-OFDMA

3GPP, http://www.3gpp.org. (Acce ss Date July20,2012) and SC-FDMA base d Radio Access, John Wiley & Sons

Ali, Salman, and Zeeshan, Muhammad, "A Utility Base d Ltd, 2009.

Resource Allocation Scheme with De lay Sche dule r for Ikuno, Jose p Colom, Wrulich, Martin, and Rupp, Markus,

LTE Service-Class Support," IEEE Wireless Communications “System Le ve l Simulation of LTE ne tworks,” IEEE

and Ne tworking Confere nce (WCNC), pp. 1450-1455, Ve hicular Technology Confere nce (VTC ), pp. 1 - 5, May

April 2012. 2010.

Andre ws, Jeffrey G., Ghosh, Arunabha, and Muhame d, Rais, Iosif, Oana, and Banica, Ion, “On the Analysis of Packe t

Fundame ntals of WiMAX – Unde rstanding Broadband Sche duling in Downlink 3GPP LTE System," The Fourth

Wire le ss Ne tworking, Pe arson Education, Inc., 2007. International Confere nce on Communication The ory,

Capozzi, F., Piro, G., Grie co, L. A., Boggia, G., and Camarda, Re liability, and Quality of Se rvice ", pp. 99 – 102, Oct.

P., “Do wnlink Packe t Sche duling in LTE Ce llular 2011.

Ne tworks: Key Design Issues and a Survey,” IEEE IOSIF, Oana, and BĂNICĂ, Ion, “Pe rformance Analysis OF
Communications Surveys & Tutorials, Vol. 15, pp. 678 - Downlink LTE USING System Le ve l Simulator,” U.P.B.
700, June 2012. Sci. Bull., Se rie s C, Vol. 75, Jan. 2013.
Chadchan, S. M., and Akki, C. B., "A Fair Downlink Kawser, Mohammad T., Farid, Hasib M. A. B., Hasin,
Sche duling Algorithm for 3GPP LTE Ne tworks," I. J. Abduhu R., Sadik, Adil M. J., and Razu, Ibrahim K.,
Compute r Ne twork and Information Security, pp. 34 - 41, “Pe rformance Comparison be tween Round Robin and
May 2013. Proportional Fair Sche duling Me thods for LTE,”
Chadchan, S. M., and Akki, C. B., “3GPP LTE/SAE: An International Journal of Information and Electronics
Ove rvie w,” International Journal of Compute r and Engine e ring, Vol. 2, No. 5, pp. 678 – 681, Se pte mber 2012.
Ele ctrical Enginee ring, Vol. 2, No. 5, pp 806-814, October, Ke la, P., Puttone n, J., Kole hmaine n, N., Ristanie mi, T.,
2010. He nttone n, T., and Moisio, M., “Dynamic Packe t

11
www.mc-journal.org Mobile Computing Vol. 3 Iss. 1, Fe bruary 2014

Sche duling Performance in UTRA Long Te rm Evolution Schwarz, Stefan, Me hlführe r, Christian and Rupp, Markus,
Downlink,” in Proc. of International Symposium on "Throughput Maximizing Multiuse r Sche duling with
Wire less Pe rvasive Comput.,, Santorini, Greece , May Adjustable Fairness," IEEE Inte rnational Confere nce on
2008, pp. 308 –313. Communications (ICC), pp. 1 - 5, June 2011.
Khan, Nabee l, Martini, Maria G., Bharucha, Zubin, and Auer, Sun, Zhishui, Yin, Changchuan, and Yue , Guangxin,
Gunthe r, “Opportunistic Packe t Loss Fair Sche duling for “Re duce d-Comple xity Proportional Fair Sche duling for
De lay-Sensitive Applications ove r LTE Systems,” IEEE OFDMA Systems,” IEEE Inte rnational Confere nce on
Wire less Communications and Ne tworking Confere nce Communications, Circuits and Systems Procee dings, Vol.
(WCNC), pp. 1456 – 1461, April 2012. 2, pp. 1221 – 1225, June 2006.
Larmo, A., Lindström, M., Me yer, M., Pe lle tier, G., Torsner, Talevski, De jan, and Gavrilovska, Liljana, “Nove l
J., and Wie mann, H., "The LTE Link-Layer Design," IEEE Sche duling Algorithms for LTE Downlink Transmission,”
Communications Magazine , Vol. 47, pp. 52-59, Apr il TELFOR Journal, Vol. 4, No. 1, 2012.
2009. Tran, Sang V., and Eltawil, Ahme d M., "Optimize d
LI, Xiaowe i, LI, Bingbing, LAN, Bing, HUANG, Min, and YU, Sche duling Algorithm for LTE Do wnlink System," IEEE
Guanghui, “Adaptive PF Sche duling Algorithm in LTE Wire less Communications and Ne tworking Confere nce
Ce llular System,” IEEE International Confere nce on (WCNC), pp. 1472 - 1476, April 2012.
Information and Communication Technology Wang, Anchun, Xiao, Liang, Zhou, Shidong, Xu, Xiiin, and
Conve rgence (ICTC), Nov. 2010, pp. 501 - 504. Yao, Yan, “Dynamic Resource Manage me nt in The
Nonchev, Stanislav, and Valkama, Mikko, “A Ne w Fairness- Fourth Ge neration Wire less Systems,” Inte rnational
Orie nte d Packe t Sche duling Sche me with Re duce d Confe rence on Communication Te chnology Procee dings
Channe l Fee dback for OFDMA Packe t Radio Systems,” (ICCT), Vol. 2, pp. 1095 - 1098, April 2003.
Int. J. Communications, Ne twork and System Scie nces We nge rter, C., Ohlhorst, J., and von Elbwart, A. G. E.,
(IJCNS), pp. 608 - 618, June 2009. “Fairness and Throughput Analysis for Gene ralize d
Nonchev, Stanislav, and Valkama, Mikko, “Efficie nt Packe t Proportional Fair Fre que ncy Sche duling in OFDMA,”
Sche duling Schemes with Built-in Fairness Control for IEEE Ve hicular Technology Confe re nce (VTC), Vol. 3,
Multi-ante nna Packe t Radio Syste ms,” Inte rnational May 2005, pp. 1903 - 1907.
Journal on Advances in Ne tworks and Services, vol. 2, Wu, Shih-Jung, and Chu, Liou, “A Nove l Packe t Sche duling
No. 2&3, pp. 182 - 194, 2009. Scheme for Downlink LTE System,” IEEE Seve nth
Parruca, Donald, Grysla, Marius, Gortzen, Simon, and Gross, International Confere nce on Inte llige nt Information
James, “Analytical Mode l of Proportional Fair Hiding and Multime dia Signal Processing (IIH-MSP), pp.
Sche duling in Inte rfere nce -limite d OFDMA/LTE 25 - 28, Oct. 2011.
Ne tworks,” Submitte d to The IEEE for possible Xie , Tian, and Hui, Xiaowe i, “LTE System Sche duling
publication, Submitte d on 7 Mar 2013. Algorithm base d on Utility Function,” IEEE, Nov. 2011,
Piro, Giuse ppe , Grie co, Luigi Alfre do, Boggia, Ge nnaro, pp. 370 - 373.
Fortuna, Rosse lla, and Camarda, Pie tro, “Two-Le ve l Yaacoub, Elias, “Pe rformance Study of the Imple mentation
Downlink Sche duling for Real-Time Multime dia Se rvices of Green Communications in LTE Ne tworks,” IEEE
in LTE Ne tworks,” IEEE Transaction on Multime dia, Vol. International Confe rence on Te lecommunications (ICT),
13, No. 5, pp. 1052 – 1065, Oct. 2011. pp. 1 – 5, April 2012.

12

View publication stats