Adaptive Radio Resource Allocation for Downlink
OFDMA/SDMA Systems
Chun-Fan Tsai
∗, Chung-Ju Chang
∗, Fang-Ching Ren
†, and Chih-Ming Yen
∗∗Department of Communication Engineering
National Chiao Tung University Hsinchu, Taiwan 300 Email: [email protected]
†Information and Communication Research Laboratories
Industrial Technology Research Institute Hsinchu, Taiwan 300
Abstract—The paper proposes an adaptive radio resource allocation (ARRA) algorithm for downlink OFDMA/SDMA sys-tems with goals to achieve quality of service (QoS) satisfied and throughput maximized. The ARRA algorithm considers multiple service classes of multimedia traffic and their diverse QoS requirements. It contains two parts, a dynamic priority adjustment scheme and a priority-based greedy (PBG) algorithm. The dynamic priority adjustment scheme gives high priority to urgent users and dynamically adjust the value frame by frame based on users’ QoS requirement and queue occupancy. The PBG algorithm allocates the radio resource iteratively according to a cost value to maximize the system throughput while allocating enough resource to high-priority users. Simulation results show that the ARRA algorithm outperforms the conventional algo-rithms in terms of system throughput under the satisfaction of QoS requirements.
I. INTRODUCTION
Orthogonal frequency division multiple access combined with space division multiple access (OFDMA/SDMA) can be an effective approach to support high-speed wireless com-munications. The OFDMA is based on OFDM (orthogonal frequency division multiplexing) and inherits its superiority of mitigating multipath fading and maximizing spectral effi-ciency. The SDMA with beamforming technique to multiplex multiple users on the same subchannel for increasing the sys-tem throughput. For a multiuser OFDMA syssys-tem, the syssys-tem data rate was maximized when each subcarrier was assigned to the user with the best channel gain [1]. However, when SDMA is enabled in an OFDMA system, the data rate of the system is maximized while the optimal set of cochannel users is selected for each subcarrier. When channel state information (CSI) is available at base station, a sophisticated radio resource alloca-tion (RRA) scheme is needed for OFDMA/SDMA systems to maximize system throughput by exploiting system diversity.
In a modern wireless system that supports multimedia traf-fic, quality-of-service (QoS) guarantee should be a design con-sideration for RRA algorithms. An optimal resource allocation algorithm for OFDMA system was proposed to minimize the total transmission power consumption under the satisfaction of QoS requirement in [2]. The generalized processor sharing (GPS) scheduling was extended to the OFDM systems in
[3]. The resource allocation for OFDMA/SDMA system was first investigated in [4] and [5]. However, in [6], the authors challenged the performance of [4] and [5] and proposed an optimal solution for maximizing information capacity. Also, practical bit loading schemes were proposed in [7], where a heuristic approach was taken to reduce the high complexity of the RRA algorithm in OFDMA/SDMA systems.
From these previous works, three observations can be in-duced. Firstly, all above schemes can be considered as fixed-priority schemes. The resource is either allocated to guarantee a fixed number of transmission bits or assigned according to predefined weights. Since the required resource is fixed in each OFDMA symbol, the time diversity is not well exploited and system throughput is degraded. Secondly, only bit error rate (BER) and/or minimum transmission rate were considered as QoS requirements in previous RRA algorithms. However, with the present of multimedia traffic, the delay requirement should also be included. Finally, most of researches assumed that a subcarrier is used as the basic allocation unit and each user always has data in its buffer. However, a subcarrier-based allocation is difficult to realize due to its high signaling overhead. Noticeably, the basic allocation unit in a practical OFDMA system (e.g. IEEE 802.16 [8]) is a subchannel, which is a set of subcarriers. Also, in realistic environments providing various service types, the traffic models should be taken into account in the design of RRA algorithms.
The paper proposes an adaptive radio resource allocation (ARRA) algorithm for downlink OFDMA/SDMA systems with an objective to maximize the system throughput and to guarantee the QoS of multimedia traffic. The ARRA algorithm is composed of two parts. The first part is an dynamic priority adjustment scheme, where priorities of users are dynamically adjusted frame by frame and the required resource of each user varies with time. By giving high priority to urgent users, it is believed that the ARRA algorithm can attain the tradeoff between throughput and QoS requirement better than the schemes with fixed priority. The second part of the ARRA algorithm is a low-complexity resource allocation scheme, called priority based greedy (PBG) algorithm, to allocate resource based on the priority obtained in the first
Spatial Processing (Beamforming) Subchannel Allocation Adaptive Modulator Data stream of user 1 Data stream of user 2 Data stream of user K Adaptive Modulator Adaptive Modulator OFDM Transmitter OFDM Transmitter OFDM Transmitter Antenna Q Antenna 2 Antenna 1
Adaptive Radio Resource Allocation (ARRA) Algorithm 1 p 2 p Q p
Channel State Information
QoS Requirements Power
Allocation Modulation Order
Assignment
Fig. 1. Transmission structure of the OFDMA/SDMA System
part and the CSI of users. Simulation results show that the ARRA algorithm achieves the system throughput higher than conventional algorithms, under the QoS requirements.
II. SYSTEMMODEL
The OFDMA/SDMA system is assumed to support three classes of service, real-time (RT), non-real-time (NRT), and best effort (BE), which are with different QoS requirements. For RT services, the QoS requirements consider BER, maxi-mum delay tolerance, and maximaxi-mum allowable dropping ratio. For NRT services, the QoS requirements are BER and mini-mum required transmission rate. For BE services, only BER is included in the QoS requirement. Each mobile station (user) belongs to one kind of service classes, and a traffic model is associated with the user. Denote these QoS requirements of BER, minimum required transmission rate, maximum delay tolerance, and maximum allowable dropping ratio by BER∗k,
R∗k, D∗k, and PD,k∗ respectively. The system provides one
individual queue for each downlink user at the base station. Packets of RT services will be dropped if the delay of packets exceeds the maximum delay tolerance, while packets of NRT services or BE services are allowed to be queued without being dropped if buffer occupancy is not overflowed.
The architecture of the downlink OFDMA/SDMA sys-tem with the ARRA algorithm is shown in Fig. 1, where data streams for K single-antenna mobile stations will be
transmitted from the base station which is equipped with
N subchannels and Q transmit antennas. A set of OFDM
subcarriers forms an OFDMA subchannel, which is the basic unit for resource allocation and adaptive modulation in a practical OFDMA system [8]. A subchannel is assumed to have continuousb subcarriers since the grouping of continuous
subcarriers results in highest multiuser diversity [9]. The time axis is divided into frames with fixed length, and each frame includesL OFDMA symbols for downlink transmission. Three
possible modulation scheme QPSK, 16-QAM, and 64-QAM
are using in this paper and let q = 2 × b be the number
of transmission bits with the basic QPSK modulation over
b subcarriers in one subchannel. The ARRA algorithm is
executed at the beginning of every frame to properly allocate radio resource to all users according to their queue state, CSI,
and QoS requirements. The radio resource allocation includes the subchannel allocation, modulation order assignment, power allocation, and beamforming control.
For the system under consideration, let Ψnbe the set of the subcarriers in subchannel n and K()n be the set of users that are multiplexed on subchanneln for the th OFDMA symbol.
The transmit symbol vector in subcarrier i of subchannel n
for theth OFDMA symbol, denoted by S()i , is given by S()i =
k∈K() n
ξk,i()d()k,iw()k,i, i ∈ Ψn, (1)
whereξ()k,i is the allocated power,d()k,i is the data symbol, and
wk,i() is aQ × 1 beamforming vector, for user k at subcarrier
i at the th OFDMA symbol. Notice that a normalized QAM
modulation is used such that the data symbol has unitary mean energy.
A perfect CSI estimation for each user is assumed and the channel is fixed within a frame duration. Let hk,i be a 1 × Q vector denoting the frequency domain channel gain from base station to user k on subcarrier i. For simplicity
and acceptable performance, the zero-force (ZF) transmit beamforming scheme [6], [7] is used. The cochannel users are orthogonal in space domain while ZF transmit beamforming is adopted. Therefore, the received signal of userk in subcarrier
i for the th OFDMA symbol, denoted by Yk,i(), is given by,
Yk,i()= hk,iw()k,i
ξk,i()d()k,i+ Zk,i(). (2)
whereZk,i()is the thermal noise on userk in subcarrier i and is
assumed to be complex Guassian with zero mean and variance
σ2. From the above equation, the received SNR can be written as,
SNR()k,i =ξ
()
k,ihk,iw()k,i 2
σ2 . (3)
Note that the received SNR is affected by the beamforming vector. If the users with high spacial correlation are selected, the term hk,iw()k,i, will be small and a poor received SNR will be resulted. A scheduler should select the users with low spacial correlation into the same subchannel.
If the cochannel user set, K()n , is determined, the beam-forming vectors of the users in subchanneln can be calculated
by using the formulation of ZF beamforming. Then, it can be seen form (3) that the received SNR of user k depends on
the allocated power, ξk,i(). To maintain the BER performance, the allocated power to user k is set to the value such that
the received SNR is equal to the minimum required SNR of
userk, which can be obtained from BER∗k and the modulation
scheme of userk. If user k adopts M -QAM modulation, the
minimum required SNR,SNR∗k, is given by [10],
SNR∗
k= −ln(5 BER
∗ k)
According, from (3) and (4), the allocated power, ξk,i(), can be obtained by, ξk,i()= −ln(5 BER ∗ k)(M − 1) 1.5 σ2 hk,iw()k,i 2. (5)
Thus the power allocated to user k on subchannel n for the
th OFDMA symbol, denoted by p()k,n, can be calculated as
p()k,n = i∈Ψnξk,i(). In other words, the power allocated to a user should be sufficiently enough to guarantee the BER requirement if the user is selected by the ARRA algorithm.
III. ADAPTIVERADIORESOURCEALLOCATION
Define x()k,n as the assignment variable for indicating
the transmission state of user k on subchannel n at the
th OFDMA symbol, where x()k,n = 0, 1, 2, or 3 means
that no transmission, transmit using QPSK modulation, transmit using 16-QAM, and transmit using 64-QAM,
respectively. Denote the assignment vector x() ≡
x()1,1, · · · , x()1,N, · · · , x()k,1, · · · , x()k,N, · · · , x()K,1, · · · , x()K,N
T
the solution of the ARRA algorithm for the th OFDMA
symbol. Then, the allocated transmission bits to userk in this
frame, denoted by Rk, and the cochannel sets, K()n , can be
obtained from the assignment vectors. Hence, if needed,K()n
and x() will be denoted by Rk(x(1)· · · x()) and K()n (x())
in the following. Also, p()k,n can also be considered as a function of BER∗k andx().
Four constraints are considered on the design of the ARRA algorithm. The first one is the subchannel allocation
con-straint, which means that a subchannel can be allocated toQ
users at most in order to make the ZF beamforming realizable. The constraint is expressed as |K()n (x())| ≤ Q. The second one is the total system power constraint, which represents that the total power allocation for downlink data transmission have a limitation. DenotePT the total power, the constraint is
writ-ten asNn=1Kk=1pk,n()(BER∗k, x()) ≤ PT. For transmission
efficiency, the third constraint, buffer occupation constraint, is added. The allocated bits to a user should not larger than the its buffer occupancy, which can be expressed asRk≤RBk/q
·q,
where RBk is the buffer length of userk.
For further satisfying the QoS requirement of the RT users and NRT users, the last constraint, QoS fulfillment constraint, is included. First, a priority value is set for each user at each frame according to its QoS requirements and queue occupancy. We define the priority value of user k, denoted by Rk, as
the minimum number of transmission bits required at current frame in order to fulfill the user’s QoS Requirements. Thus, theRkshould have the QoS fulfillment constraint expressed as
Rk≥ Rk. At least Rk bits should be allocated to userk at the
current frame to guarantee its QoS requirements. Noticeably, the priority value of a user is dynamically adjusted frame by frame.
Therefore, the ARRA algorithm is formulated as an
opti-mization problem given by,
(x∗(1)· · · x∗(L)) = arg max x(1)···x(L) K k=1 Rk(x(1)· · · x(L))
subject to the following constraints:
|K() n (x())| ≤ Q, ∀n, , N n=1 K k=1 p()k,n(BER∗k, x()) ≤ PT, ∀, Rk≤RBk/q · q, ∀k, Rk≥ Rk, ∀k. (6)
In this formulation, the system throughput is maximized under the four constraints. The optimization problem (6) can be easily solved by an integer programming method [11]. How-ever, the complexity of the integer programming method grows exponentially with the number of users and is unacceptable in real applications. Hence, the proposed ARRA algorithm adopts a reduced-complexity approach based on greedy approach [11]. The proposed ARRA algorithm contains two parts to find the solution in (6), an dynamic priority adjustment scheme and a priority-based greedy (PBG) algorithm. The details are described in the following.
A. Dynamic Priority Adjustment Scheme
We here introduce a time-to-expiration (TTE) value, indi-cating the urgency degree of a user at the current frame. For
userk, denote the TTE value and the number of residual bits
of the head-of-line (HOL) packet byVk andBk, respectively.
The smaller theVk is, the more the degree of urgency of userk
would be. For users with RT service class, theVk is intuitively
given by,
Vk = D∗k− Dk, (7)
where Dk is the time duration from the arrival of the HOL
packet of userk to the current frame, and the unit of both Dk
andD∗k is in frames. For users in NRT service class, theVk
is given by, Vk = Bk + Bk R∗k − D k , (8)
whereDk is the the time duration while there is data buffered in the queue of userk before the current frame, Bk is the total transmission bits of user k in Dk, and R∗k is the minimum required transmission rate in a unit of bits per frame. The
derivation of Vk in (8) of NRT user k comes from the
inequality (Bk+ Bk)/(Vk + Dk) ≥ R∗k, which means that
the average rate should be greater than the minimum required transmission rate. Finally, for users in BE service class, the
Vk is intuitively set to be infinity.
GivenVkandBkof userk, its priority value, Rk, is defined
as, Rk= 0, ifVk= ∞ Bk q · q, ifVk≤ Vth maxBk Vk·q − ln(Vk) , 0 · q, elsewise, (9)
whereVthis a threshold forVk. IfVk= ∞, then it is intuitive
to set Rk as zero. IfVk is below the threshold Vth, it means that the degree of urgency of userk is very high in a fashion
that the userk should complete its transmission in this current
frame, i.e. Rk =
Bk
q
· q. Otherwise, the design of Rk
is based on the average required transmit bits in remaining frames, Bk/Vk, added with a negative bias (− ln(Vk)). The
negative bias reduces the priority of the delay-tolerable users, so that the system can give the transmission opportunity to other high-urgent users. Note that a user with low priority could still be served by the base station if the channel quality of the user is good and other users with higher priority have been already served. Hence, the delay-tolerable users can take the advantage of time diversity by transmitting only when its channel is good. As for the threshold value Vth, it could be
set to one if resource is always enough to satisfy Rk ≥ Rk.
However, since the user might be in cell boundary, the Vth
could be set to a value greater than one to guarantee the QoS requirement earlier. In the later section of simulation, theVth
is set to three.
B. PBG Algorithm
The basic principle of the PBG algorithm is that every successive step is taken to minimize an immediate cost. The immediate cost is the increment of power of increasing one modulation order for a user on one subchannel. Denote Ck,n()
the cost function of userk on subchannel n at the th OFDMA
symbol. If 0≤ x()k,n≤ 2, the cost is calculated as,
Ck,n() =
k∈K()n (x+())
p()k,n(BER∗k, x+()) − p()k,n(BER∗k, x()),
(10) where x+() is the assignment vector after the modulation
of user k on subchannel n is increased by one given the
current x(). The x+() will be the same as x() except
x+()k,n = x()k,n+ 1. Otherwise, Ck,n() is set to infinity since the maximum modulation order is reached and the modulation order can not be increased further. Increasing of modulation order from zero to one means adding a new user to a sub-channel, which requires recalculation of beamforming vectors. The definition of the cost value also includes the increasing power for maintaining the same modulation order for the users that already in the subchannel. Hence the spatial correlation between the new user and the users that are already in the subchannel is also measured by the cost function.
The PBG algorithm is initialized as follows. The assignment variables are set to zero, and each user is given an instanta-neous priority. Denoteβk the instantaneous priority of userk,
and it is set as the priority from dynamic priority adjustment scheme. Denote P() the current used power and Nfree() the set of free subchannels for theth OFDMA symbol. They are
initialized as P() = 0 and Nfree() = {n|1 ≤ n ≤ N}, ∀.
The PBG algorithm then sequentially allocates resource for each OFDM symbol used for downlink transmission in the current frame. In each symbol, two functions are performed,
function Allocation-for-one-symbol and function Extend. The PBG algorithm is depicted in the following pseudocode.
• PBG Algorithm
Setx()= 0, ∀ and βk= Rk, ∀k.
SetP()= 0 and Nfree()= {n|1 ≤ n ≤ N}, ∀.
for = 1 : L do
Execute function Allocation-for-one-symbol. Execute function Extend.
end for
In function Allocation-for-one-symbol, an iterative algo-rithm is executed for resource allocation in symbol . A
candidate user set, denoted by Ω, is constructed and an optimal pair of user and subchannel, (k∗, n∗), is selected, in every
iteration. The Ω contains the backlogged users with highest instantaneous priority and the (k∗, n∗) is selected from the
users in the candidate user set and the free subchannels such that the cost value, Ck()∗,n∗, is minimum. If the power budget
in theth OFDMA symbol is still sufficient for increasing the
modulation order for user k∗ on subchannel n∗, then some states are updated as follows. The modulation order of the selected user on the selected subchannel is increased by one, i.e. x()k∗,n∗ = x()k∗,n∗ + 1. Additional q bits are allocated
to the selected user in each iteration, and thus the queue length of user k∗, RBk, is decreased by q. Used power for
the th OFDMA symbol, P(), is increased by the minimum
cost. For fairness issue, the instantaneous priority of the user
k∗ is decreased by q until the priority become zero, i.e. βk∗ = max(βk∗ − q, 0). Thus, low-priority users can still
have opportunity to be transmitted. The solution of resource allocation in the th OFDMA symbol is given in vector x()
while this function is terminated.
In function Extend, the same allocation could be extended over several OFDMA symbols to form a time burst trans-mission. This function is to reduce the signaling overhead of the system and the complexity of the PBG algorithm. If
subchannel n the th OFDMA symbol has been allocated
to a specified group of users, according to the result of function Allocation-for-one-symbol, then we can allocate the subchanneln in the +1th OFDMA symbol to the same group
of users. The same operation, which is called extension in this paper, can be done for symbol + 2, + 3, and so on. The
extension would be performed as long as the current queue occupancy for each user in the specified group is not empty. The assignment variables, instantaneous priority, queue length, used power, and the set of free subchannels are updated if the extension is performed.
IV. SIMULATIONRESULTS ANDDISCUSSION
The downlink OFDMA/SDMA system environments are configured according to the IEEE 802.16 standard [8], where parameters are listed in Table I. The path loss model is modeled as 128.1 + 37.6 log R dB, where R is the distance
between the base station and the user in kilometers [12]. The log-normal shadowing is assumed with zero mean and standard
TABLE I
OFDMA/SDMA SYSTEMPARAMETERS
Parameters Values
Cell size 1600m
Number of antenna at base station (Q) 3
Frame duration 2ms
System bandwidth 5 MHz
FFT size 512
Number of data subcarriers 384
Number of subchannels (N) 8
Number of data subcarriers per subchannel (b) 48
Number of OFDMA symbol for downlink 8
transmission per frame (L)
Power allocation to data transmission (PT) 43.10 dBm
Thermal noise density -174 dBm/Hz
TABLE II
THEQOS REQUIREMENT OF EACH TRAFFIC TYPE
Required Maximum Maximum Minimum Required
BER Delay Allowable Transmission
Tolerance Dropping Ratio Rate
Voice 10−3 40ms 1%
-Video 10−4 10ms 1%
-HTTP 10−6 - - - 100 kbps
FTP 10−6 - - -
-deviation of 8 dB. The multipath channel for each antenna is modeled as six taps of Rayleigh-faded paths. Four kinds of traffic types are assumed in the system, voice traffic [13] of RT service, streaming video traffic of RT service [12], HTTP traffic of NRT service [12], and FTP traffic of BE service [12]. Note that channel coding is not used in the simulation for reducing simulation time. The number of users is increased from 40 to 600, and each traffic type has the same number of users. The traffic load is defined as the ratio of the total average arrival rate of all users over the system maximum transmission rate, which could be achieved when Q users are multiplexed
for each subchannel and the highest modulation order is used. Table II lists the QoS requirements of each traffic type.
Three conventional RRA algorithms will be considered to compare with the proposed ARRA scheme. The first one is linkgain-based resource allocation (LBRA) [1], where resource is allocated to users according to users’ CSI. For each subchan-nel, the first Q users with best channel quality are selected.
The second one is multi-antenna multi-user maximum sum rate (MMSR) [7], which contains a user clustering procedure and a bit-removing algorithm. The former selects the set of cochannel users while the later determine the modulation orders of the select users. The last one is truncated generalized processor sharing (TGPS) [3], where resource is allocated to users based on predefined weights of all users. The predefined weight is set to 10, 5, and 1 for RT, NRT, and BE services, respectively.
Fig. 2 shows the system throughput versus the traffic load. It can be found that the system throughput of the ARRA scheme performs the best. The reasons are: the ARRA algorithm improves the system throughput by taking multiuser diversity and space domain correlation between users in its design. The system throughput of the MMSR scheme is near to
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 5 10 15 20 25 Traffic Load System Throughput (Mbps) ARRA LBRA MMSR TGPS
Fig. 2. System Throughput
0 0.2 0.4 0.6 0.8 1 0 0.05 0.1 0.15 0.2 0.25
Dropping Ratio Requirement
Traffic Load (a)
Voice Packet Dropping Ratio
ARRA LBRA MMSR TGPS 0 0.2 0.4 0.6 0.8 1 0 0.05 0.1 0.15 0.2 0.25
Dropping Ratio Requirement
Traffic Load (b)
Video Packet Dropping Ratio
ARRA LBRA MMSR TGPS
Fig. 3. (a) Packet Dropping Ratio of Voice Users (b) Packet Dropping Ratio of Video Users
that of the ARRA scheme because both of this two schemes takes throughput maximization as a design objective. The system throughput of LBRA scheme is less than that of the ARRA scheme for the reason that the optimal user grouping in space domain is not considered in LBRA algorithm. The system throughput of TGPS algorithm is smallest in the four algorithms; it is because the TGPS uses simplified algorithm for subchannel allocation and the multiuser diversity is not well exploited.
Figs. 3 (a) and 3 (b) depict the packet dropping ratio of voice users and the packet dropping ratio of video users, respectively. It can be seen that the voice/video packet dropping ratios of the ARRA algorithm are almost zero until the traffic load is greater than 0.8, while those of the other algorithms increase
rapidly with the traffic load. The reason is that the LBRA or MMSR algorithm does not consider the QoS requirement of maximum delay tolerance for the RT users. As for the TGPS algorithm, since its maximum capacity is too small, its dropping ratio became large at high traffic load even though the TGPS algorithm gives large weights to RT users. On
0 0.2 0.4 0.6 0.8 1 0 200 400 600 800 1000 1200 1400 1600
Minimum Rate Requirement
Traffic Load (a)
Average Transmission Rate (Kbps)
ARRA LBRA MMSR TGPS 0 0.2 0.4 0.6 0.8 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Traffic Load (b)
Guaranteed Ratio of HTTP Users
ARRA LBRA MMSR TGPS
Fig. 4. (a) Average Transmission Rate of HTTP Users (b) Guaranteed Ratio
of HTTP Users
the other hand, the ARRA algorithm considers the RT user with large packet delay as urgent users and gives the users high priority. Since resource is first provided for high-priority users, the delay requirement of the RT user is satisfied and the dropping ratio is the smallest.
Figs. 4 (a) and 4 (b) illustrate the average transmission rate of HTTP users and the guaranteed ratio of HTTP users, respectively. The guaranteed ratio of NRT users is defined as the ratio of the number of the QoS-satisfied HTTP users over total HTTP users. For the ARRA algorithm, the average transmission rate decreases as the traffic load increases, but the minimum required transmission rate for NRT users is guaranteed. The transmission rate is guaranteed by giving high priority to the NRT users whenever their transmission rate is going to be lower than minimum required transmission rate. For the same reason, the guaranteed ratio of HTTP users is almost 100% in the ARRA algorithm when traffic load is low and still larger than 95% when the traffic load is 0.9. Although
the average transmission rate of the MMSR or TGPS algorithm is higher than that in the ARRA algorithm, the guaranteed ratio drops earlier than the ARRA algorithm. For example, when traffic load is 0.8 the guaranteed ratio of the ARRA algorithm
is 99% while that of TGPS algorithm and MMSR algorithm is only 70% and 40%, respectively. The LBRA algorithm has lowest guaranteed ratio of HTTP users since it only guarantees the transmission rate of the users with good channel quality.
V. CONCLUSIONS
In this paper, an adaptive radio resource allocation (ARRA) algorithm is proposed for downlink OFDMA/SDMA systems. The proposed ARRA algorithm contains a dynamic priority adjustment scheme to dynamically adjust the priority of users frame by frame, where NRT users with lower average trans-mission rate which is near the minimum required transtrans-mission rate and RT users which are with larger packet delay can be promoted to higher priority to obtain the enough resource earlier. It also consists of a PBG algorithm to efficiently
allocate the resource based on a cost value. Simulation results show that the ARRA algorithm outperforms the conventional algorithms in term of system throughput, under the satisfaction of QoS requirements.
ACKNOWLEDGMENT
This work was supported by National Science Council, Taiwan, under contract number NSC 95-2752-E-009-014-PAE and Ministry of Education, Taiwan under Grants 95W803C.
REFERENCES
[1] J. Jang and K. B. Lee, “Transmit power adaptation for multiuser OFDM system,” IEEE J. Select. Areas Commun., vol. 21, pp. 171–178, Feb. 2003. [2] C. Y. Wong, R. S. Cheng, K. B. Letaief, and R. D. Murch, “Multiuser OFDM with adaptive subcarrier, bit, and power allocation,” IEEE J.
Select. Areas Commun., vol. 17, pp. 1747–1758, Oct. 1999.
[3] J. Cai, X. Shen, and J. W. Mark, “Downlink resource management for packet transmission in OFDM wireless communication systems,” IEEE
Trans. Wireless Commun., vol. 4, pp. 2726–2737, July 2005.
[4] S. Thoen, L. V. der Perre, M. Engels, and H. D. Man, “Adaptive loading for OFDM/SDMA-based wireless networks,” IEEE Trans. Commun., vol. 50, pp. 1798–1810, Nov. 2002.
[5] I. Koutsopoulos and L. Tassiulas, “Adaptive resource allocation in SDMA-based wireless broadband networks with OFDM signaling,” in Proc.
IEEE Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies (INFOCOM 2002), vol. 3, June 2002, pp.
1376–1385.
[6] Y. M. Tsang and R. S. Cheng, “Optimal resource allocation in SDMA / muilti-input-single-output / OFDM systems under QoS and power constraints,” in Proc. IEEE Wireless Communications and Networking
Conference (WCNC-2004), vol. 3, Mar. 2004, pp. 1595 – 1600.
[7] D. Bartolome, A. I. Perez-Neira, and C. Ibars, “Practical bit loading schemes for multi-antenna multi-user wireless OFDM systems,” in Proc.
Thirty-Eighth Asilomar Conference on Signals, Systems and Computers,
vol. 1, Nov. 2004, pp. 1030 – 1034.
[8] Local and metropolitan area networks - Part 16: Air Interface for Fixed
Broadband Wireless Access Systems, IEEE Standard Std. 802.16-2004.
[9] M. Shen, G. Li, and H. Liu, “Effective of traffic channel configuration on the orthogonal frequency division multiple access downlink performance,”
IEEE Trans. Wireless Commun., vol. 4, pp. 1901–1913, July 2005.
[10] A. J. Goldsmith and S.-G. Chua, “Variable-rate variable-power MQAM for fading channels,” IEEE Trans. Commun., vol. 45, pp. 1218–1230, Oct. 1997.
[11] K. G. Murty, Operations Research. Prentice Hall, 1995.
[12] 3GPP TR 25.892, “Feasibility study for OFDM for UTRAN enhance-ment,” 3rd Generation Partnership Project, Tech. Rep., 2004-06. [13] Universal Mobile Telecommunications System, Selection procedures for
the choice of radio transmission technologies of the UMTS, UMTS Std.
