Fair and Energy-Efficient Resource Allocation Optimizationin Wireless Networks

다국어 초록 (Multilingual Abstract)

The use of network utility maximization (NUM) paradigm for the overall performance of communication
networks is to decompose the whole problem into sub-problems at various network layers, i.e., cross-layer
design. Among the problems addressed in cross-layer designs, congestion control has been regarded as
the key issue, since appropriate congestion control schemes can ensure network stability and acceptable
performance. For a part of the Ph.D. studies, we survey the state of the art cross-layer congestion control
in wireless networks and propose two congestion control schemes in multipath lossy wireless networks and
complex communications systems. This part is however not covered in this thesis and can be found in our
articles in the section Publications.
The rapid expansion of wireless communication networks drives the research community to design
wireless networks with higher spectral efficiency and energy efficiency. Besides, fairness among mobile
users in wireless networks is of critical importance. To satisfy QoS requirements and guarantee fairness in
next-generation networks, many of technology and network architecture evolution have been proposed, for
example, heterogeneous networks (HetNets), device-to-device (D2D) communication, massive multipleinput-
multiple-output (massive-MIMO), and non-orthogonal multiple access (NOMA). This thesis considers
three fair and energy-efficient resource allocation problems in such kinds of wireless networks. In
particular, various power control schemes are proposed for interference management in HetNets, for the tradeoff between spectral efficiency and energy efficiency in spectrum-sharing wireless networks, and for
fairness in NOMA systems.
Our first work considers energy-efficient power control schemes for interference management in uplink
spectrum-sharing heterogeneous networks, consisting of a higher-tier macrocell and multiple lower-tier
smallcells, where the optimization problem is formulated based on the multi-objective formulation subject
to constraints on rate outage probability and maximum tolerable interference at the macro base station. In
the first scenario, the objective function is defined as the weighted sum of the energy efficiencies and the
optimization problem is in a sum-of-ratios form, which cannot be conventionally solved by the Dinkelbachs
procedure; we develop an efficient global optimization algorithm with global linear and local quadratic rate
of convergence to solve the considered problem. To ensure fairness among individual UEs in term of energy
efficiency, we consider the max-min problem, where the objective is defined as the weighted minimum of
the energy efficiencies and a fractional programming theory and the dual decomposition method are jointly
used to solve the problem and investigate an iterative algorithm. We further discuss the global energy
efficiency problem and consider near optimal schemes. Numerical examples are provided to demonstrate
significant improvements of the proposed algorithms over existing ones.
The second work introduces a fair and energy-efficient resource allocation framework in spectrumsharing
wireless networks with quality-of-service guarantees. Consider the tradeoff between energy efficiency
and spectral efficiency, the multiobjective problem of spectral efficiency and energy efficiency is
transformed into a problem that minimizes the total power consumption and maximizes the achievable
utility, subject to power constraints and rate outage probability constraints. We then analyze the complexity
of the considered problem; particularly, the optimization problem is NP-Hard when 0 < < 1 and
= 0 and is convex for other values of the fairness index . After that, we adopt the successive convex
approximation approach to approximate and transform the NP-hard nonconvex optimization problem into
a sequence of convex programs and propose two iterative successive convex approximation (SCA) based
resource allocation algorithms. Extensive simulation results are presented to demonstrate the effectiveness and outperformance of the proposed algorithms over existing frameworks.
NOMA is now considering as a promising radio access technique for next-generation networks owing
to its offered benefits, e.g., spectral efficiency improvement. Due to the successive interference cancellation
(SIC) order at receivers, fairness among users in NOMA may not be guaranteed. Our third work focuses
on -fair resource allocation in NOMA. The complexity of the considered problem is then analyzed. In
particular, the problem is shown to be convex when 1 < 1 and = 1, NP-Hard when 0 < < 1,
and polynomial time solvable when = 0. Finally, simulation results are provided to examine effects of
the fairness degree on the system performance and verify the effectiveness of our proposed algorithms.

목차 (Table of Contents)

• Contents vi
• List of Figures ix
• 1 Introduction 1
• 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
• 1.2 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
• 1.3 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
• 2 Energy-Efficient Power Control for Uplink Spectrum-Sharing Heterogeneous Networks 10
• 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
• 2.2 Network Model and Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . 13
• 2.3 Energy-Efficient Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
• 2.4 Max-Min Energy Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
• 2.5 Discussion and Extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
• 2.5.1 Global Energy Efficiency Maximization . . . . . . . . . . . . . . . . . . . . . . . 31
• 2.5.2 Near-Optimal Energy Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
• 2.6 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
• 2.6.1 Simulation Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
• 2.6.2 Performance of the Proposed Algorithms . . . . . . . . . . . . . . . . . . . . . . 36
• 2.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
• 3 Fairness-Aware Spectral and Energy Efficiency in Spectrum-Sharing Wireless Networks 44
• 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
• 3.2 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
• 3.3 Complexity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
• 3.3.1 = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
• 3.3.2 Proportional fairness, = 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
• 3.3.3 Max-min fairness, = 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
• 3.3.4 Fairness with 1 < < 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
• 3.3.5 Fairness with 0 < < 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
• 3.4 Optimal Resource Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
• 3.4.1 SCA-based Resource Allocation with Logarithmic Approximation . . . . . . . . . 57
• 3.4.2 SCA-based Resource Allocation with D.C. Approximation . . . . . . . . . . . . . 61
• 3.5 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
• 3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
• 4 -Fair Resource Allocation in Non-Orthogonal Multiple Access Systems 76
• 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
• 4.2 System Model and Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
• 4.3 Complexity Analysis and Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
• 4.3.1 = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
• 4.3.2 Proportional fairness, = 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
• 4.3.3 Max-min fairness, = 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
• 4.3.4 Fairness with 1 < < 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
• 4.3.5 Fairness with 0 < < 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
• 4.4 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
• 4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
• 5 Conclusion and Future Work 89
• Bibliography 91
• Curriculum Vitae 100
• List of Research Achievements 101