Resource Allocation in Multiuser Multicarrier Wireless Systems

129 Pages·2007·2.08 MB·English

by  Ian C. Wong

RESOURCE ALLOCATION IN
MULTIUSER MULTICARRIER
WIRELESS SYSTEMS
RESOURCE ALLOCATION
IN MULTIUSER MULTICARRIER
WIRELESS SYSTEMS
Ian Wong
TheUniversityofTexasatAustin
DepartmentofElectricalandComputerEngineering
Austin, Texas
Brian Evans
TheUniversityofTexasatAustin
DepartmentofElectricalandComputerEngineering
Austin, Texas
IanWong BrianEvans
TheUniversityofTexasatAustin TheUniversityofTexasatAustin
DepartmentofECE DepartmentofECE
Austin,TX78712 Austin,TX78712
[email protected] [email protected]
LibraryofCongressControlNumber:2007934522
ResourceAllocationinMultiuserMulticarrierWirelessSystems
byIanWongandBrianEvans
ISBN-13:978-0-387-74944-0 e-ISBN-13:978-0-387-74945-7
Printedonacid-freepaper.
(cid:2)c2008SpringerScience+BusinessMedia,LLC.
Allrightsreserved.Thisworkmaynotbetranslatedorcopiedinwholeorinpartwithoutthewritten
permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York,
NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in
connectionwithanyformofinformationstorageandretrieval,electronicadaptation,computersoftware,
orbysimilarordissimilarmethodologynowknownorhereafterdevelopedisforbidden.
Theuseinthispublicationoftradenames,trademarks,servicemarksandsimilarterms,eveniftheyare
notidentifiedassuch,isnottobetakenasanexpressionofopinionastowhetherornottheyaresubject
toproprietaryrights.
9 8 7 6 5 4 3 2 1
springer.com
For our lovely wives, Lynn and Mouna
Preface
Next-generation broadband wireless standards, e.g. IEEE 802.16e and Third
Generation Partnership Project – Long Term Evolution (3GPP-LTE), use
Orthogonal Frequency Division Multiple Access (OFDMA) as the preferred
physicallayermultipleaccessscheme,esp.forthedownlink.Duetothelimited
resources available at the base station, e.g. bandwidth and power, intelligent
allocation of these resources to the users is crucial for delivering the best
possible quality of service (QoS) to the consumer with the least cost.
The problem of allocating time slots, subcarriers, rates, and power to the
different users in an OFDMA system has been an area of active research
in recent years. Previous research efforts in OFDMA resource allocation have
typicallyfocusedonmaximizinginstantaneousperformance,i.e.theallocation
decisions are performed for the current time instant subject to the current
resourceconstraints,whichisunabletofullyutilizethetime-varyingnatureof
thewirelesschanneltoimprovethecommunicationperformanceofthesystem.
This book focuses instead on maximizing time-averaged rates, allowing us to
exploit the temporal dimension to improve performance.
Furthermore, due to the difficult combinatorial nature of the problem,
many researchers in the past have focused on developing sub-optimal heuris-
tic algorithms. This book proposes a unified algorithmic framework based
on dual optimization techniques that have complexities that are linear in
the number of subcarriers and users, and that achieve negligible optimality
gapsinstandards-basednumericalsimulations.Adaptivealgorithmsbasedon
stochastic approximation techniques are also proposed, which are shown to
achieve similar performance with even much lower complexity.
Finally, it was assumed in previous work that perfect channel state infor-
mation (CSI) is available at the transmitter, which is quite unrealistic due to
inevitable channel estimation errors and feedback delay. This book develops
algorithmsassumingthatonlyimperfectCSIisavailable,suchthatallocation
decisions are made while explicitly considering the error statistics of the CSI.
Austin, TX Ian Wong
June 2007 Brian Evans
Contents
1 Introduction............................................... 1
1.1 Next-generation Wireless Communication Systems........... 2
1.1.1 Evolution of Cellular Standards to 3GPP-LTE ........ 3
1.1.2 Evolution of Broadband Access Standards to IEEE
802.16e .......................................... 6
1.2 Orthogonal Frequency Division Multiple Access ............. 8
1.2.1 Overview of OFDM................................ 9
1.2.2 Overview of OFDMA .............................. 10
1.3 Summary............................................... 12
1.3.1 Thesis Statement.................................. 12
1.3.2 Contributions..................................... 13
1.3.3 Organization...................................... 14
1.4 Nomenclature........................................... 14
2 Background ............................................... 17
2.1 Introduction ............................................ 17
2.2 Review of Related Work.................................. 17
2.2.1 Scheduling in Wireless Networks .................... 17
2.2.2 Multiuser Information Theory....................... 18
2.2.3 Physical Layer (PHY) Transmit Optimization......... 18
2.2.4 PHY-MAC Cross-layer Optimization................. 21
2.2.5 Comparison of Related Work ....................... 21
2.3 A New Approach to OFDMA Resource Allocation ........... 22
2.4 System Model........................................... 24
2.4.1 OFDMA Signal Model ............................. 24
2.4.2 Multiuser Statistical Fading Channel Model .......... 24
2.4.3 Optimization Variables............................. 27
2.4.4 PHY-MAC Interaction ............................. 29
2.5 Conclusion ............................................. 29
X Contents
3 Weighted-sum rate Maximization with Perfect CSI ........ 31
3.1 Introduction ............................................ 31
3.2 Continuous Rate Maximization with perfect CSI and CDI .... 31
3.2.1 Problem Formulation .............................. 31
3.2.2 Dual Optimization Framework ...................... 33
3.2.3 Numerical Evaluation of the Expected Dual .......... 36
3.2.4 Optimal Subcarrier and Power Allocation ............ 37
3.2.5 Complexity Analysis............................... 38
3.2.6 Instantaneous Weighted Sum Rate Maximization ...... 38
3.2.7 constant power Allocation .......................... 41
3.2.8 Analysis of the Duality Gap ........................ 41
3.3 Discrete Rate Maximization with perfect CSI and CDI ....... 46
3.3.1 Problem Formulation .............................. 46
3.3.2 Dual Optimization Framework ...................... 46
3.3.3 Numerical Evaluation of the Expected Dual .......... 48
3.3.4 OptimalDiscreteRate,Subcarrier,andPowerAllocation 49
3.4 Numerical Results ....................................... 50
3.4.1 Continuous Rate Allocation......................... 51
3.4.2 Discrete Rate Allocation ........................... 54
3.4.3 Complexity Comparison............................ 56
3.5 Conclusion ............................................. 57
4 Weighted-Sum Rate Maximization with Partial CSI ....... 59
4.1 Introduction ............................................ 59
4.2 Partial Channel State Information Model................... 60
4.3 Continuous Rate Maximization with Partial CSI and CDI .... 62
4.3.1 Problem Formulation .............................. 62
4.3.2 Dual Optimization Framework ...................... 63
4.3.3 Power Allocation Function Approximation............ 64
4.3.4 Optimal Subcarrier and Power Allocation ............ 65
4.3.5 Complexity Analysis............................... 66
4.4 Discrete Rate Maximization with Partial CSI and CDI ....... 66
4.4.1 Closed-form average BER function .................. 67
4.4.2 Closed-form power allocation function................ 68
4.4.3 Closed-form average rate function ................... 68
4.4.4 Problem Formulation .............................. 69
4.4.5 Dual Optimization Framework ...................... 70
4.4.6 Complexity Analysis............................... 71
4.5 Numerical Results ....................................... 72
4.5.1 Continuous Rate Case ............................. 72
4.5.2 Discrete Rate Case ................................ 72
4.5.3 Complexity Comparison............................ 76
4.6 Conclusion ............................................. 76
Contents XI
5 Rate Maximization with Proportional Rate Constraints ... 79
5.1 Introduction ............................................ 79
5.2 Proportional Rate Maximization with Perfect CSI and CDI... 80
5.2.1 Problem Formulation .............................. 80
5.2.2 Dual Optimization Framework ...................... 81
5.2.3 Computation of the Per-user Ergodic Rate............ 83
5.2.4 Complexity Analysis............................... 85
5.2.5 Extension to Discrete Rates and/or Imperfect CSI..... 85
5.3 Adaptive algorithms for Rate Maximization without CDI..... 85
5.3.1 Overview of Stochastic Approximation ............... 85
5.3.2 Stochastic Approximation Solution to the Dual Problem 86
5.3.3 Proof of Convergence .............................. 88
5.3.4 Complexity Analysis............................... 90
5.3.5 Extension to Other Formulations .................... 90
5.4 Results and Discussion ................................... 91
5.5 Conclusion ............................................. 93
6 Conclusion ................................................ 95
6.1 Summary............................................... 95
6.2 Future Work............................................ 97
6.2.1 Other Formulations................................ 97
6.2.2 MAC-PHY Cross-layer Scheduling................... 99
6.2.3 MIMO-OFDMA................................... 99
6.2.4 Multi-cell OFDMA ................................ 99
6.2.5 Multi-hop OFDMA................................100
A Derivation of the inverse function (3.10) of g (3.8) .......101
m,k
B Proof of Prop. 3.1 .........................................103
C Derivation of (3.35) ........................................105
D Derivation of the cdf (3.39) and pdf (3.40) of gd (3.38) ....107
m,k
E Derivation of (4.30) ........................................109
References.....................................................111
Index..........................................................117
List of Tables
1.1 Wireless data applications and their required data rates [1] .... 2
1.2 IEEE 802.16e OFDMA Scalability Parameters [2] ............ 8
2.1 Related work comparison.................................. 22
2.2 Notation Glossary ........................................ 25
3.1 Simulation Parameters .................................... 51
3.2 Relevant Performance Measures for the Continuous Rate
Resource Allocation Algorithms ............................ 54
3.3 Relevant Performance Measures for the Discrete Rate
Resource Allocation Algorithms ............................ 56
3.4 Comparison of the proposed ergodic and instantaneous rate
resource allocation algorithms with constant power allocation
algorithm assuming perfect CSI and CDI. ................... 57
4.1 Comparison of the proposed continuous and discrete rate
resource allocation algorithms assuming partial CSI, with the
suboptimal method of using the perfect CSI algorithms on the
imperfect CSI. ........................................... 71
4.2 Simulation Parameters .................................... 72
4.3 Relevant Performance Measures for the Proposed Resource
Allocation Algorithms .................................... 74
6.1 Comparison of proposed algorithms with previous work ....... 96
List of Figures
1.1 Cellular wireless communication system with hexagonal cells.
Different shading patterns of the cells indicate different sets of
frequency allocations. ..................................... 3
1.2 Frequency division multiple access (FDMA) with frequency
division duplexing (FDD).................................. 4
1.3 Timedivisionmultipleaccess(TDMA)withfrequencydivision
duplexing (FDD). ........................................ 5
1.4 Codedivisionmultipleaccess(CDMA)withfrequencydivision
duplexing (FDD). ........................................ 6
1.5 Orthogonal frequency division multiple access (OFDMA) with
either TDD or FDD....................................... 7
1.6 OFDM baseband spectrum, showing the broadband channel
subdivided into a multitude of narrowband subchannels........ 9
1.7 OFDM transmitter block diagram [3] ....................... 10
1.8 OFDM receiver block diagram [3] .......................... 10
1.9 OFDMA resource allocation for M users. Each user is
assumed to have statistically independent channel gains, and
are allocated a different set of subcarriers by the base station... 11
1.10 K-subcarrier OFDMA system block diagram for M users.
Each user is allocated a different set of subcarriers by the base
station. ................................................. 11
2.1 Example discrete-rate function for an uncoded system with
BER=10−3. Note that the SNR is plotted in linear and not
dB scale. ................................................ 28
3.1 Multi-level waterfilling snapshot for a 2-user, 76-subcarrier
system. ................................................. 35
3.2 Per-user and per-subcarrier dual g and the corresponding
m,k
optimal subcarrier and power allocation p∗ ................. 36
m,k