Multi-Timescale Cross-Layer Designs for Wireless Multihop Networks

다국어 초록 (Multilingual Abstract)

ABSTRACT

Multi-Timescale Cross-Layer Designs for Wireless Multihop Networks
Quoc-Viet Pham
(Advisor: Prof. Won-Joo Hwang)
Department of Information and Communication System
Graduate School, Inje University

Since the publication point of the Kelly’s paper, a numerous number of researches have been devoted to resource allocation and cross-layer designs in wired networks as well as wireless networks.
Most of cross-layer algorithms are established through the concepts of network optimization, especially, convex optimization, which has been the accumulated results of many researches
and many years. In the context of cross-layer optimization, we propose two cross-layer designs in fast-fading lossy delay-constrained wireless multihop networks.
The first cross-layer problem we study is to increase the overall utility and decrease the link delay and power consumption subject to constraints on link rate outage probability, link congestion
control, and flow rate conservation, in mobile ad hoc networks. As opposed to previous work,the rate outage probability in this work is based on exactly-closed form; therefore, the proposed
method can guarantee the globally optimal solution to the underlying problem. The non-convex formulated problem is transformed into a convex one, which is solved by exploiting the duality
technique. Numerical simulations verify that our proposal can achieve considerable benefit over the existing method.
Conventionally, cross-layer designs with same-timescale updates can work well; however,there is a difference in layers’ timescales and each layer normally operates at its corresponding
timescale when implemented in real systems. Respecting this issue and realizing the same problem as in the first work, the second cross-layer design we propose takes into account the timescale
difference among layers. By using the primal decomposition technique, the network optimization problem is decomposed into subproblems at various layers, from which the proposed algorithm
can be implemented in a distributed manner and adheres to the natural timescale difference among layers. Our simulation results show that the proposed design yields higher effective rates, consumes
less power and suffers less delay in comparison with the current alternative frameworks.
In addition, the design adheres to the natural timescale separation, then improves the convergence speed over the corresponding same-timescale method.
Keywords: Rate Control, Link Delay, Power Allocation, Lossy Links, Rayleigh-fading channels, Multi-Timescale, Cross-Layer Optimization.

목차 (Table of Contents)

• Contents v
• List of Figures viii
• Nomenclature viii
• 1 Introduction 1
• 1.1 Background and Motivation 1
• 1.2 Contributions 6
• 1.3 Thesis Outline 7
• 2 Globally Optimal Solutions for Cross-Layer Design in Fast-Fading Lossy Delay-Constrained MANETs 8
• 2.1 Introduction 8
• 2.2 Related Work 10
• 2.3 System Model 11
• 2.3.1 Network Model 11
• 2.3.2 Average Delay 12
• 2.3.3 Rate-Outage Probability and Effective Rate 13
• 2.3.4 Optimization Problem 15
• 2.4 nRENUM Distributed Algorithm 17
• 2.5 Simulation Results 22
• 2.5.1 Simulation Settings 22
• 2.5.2 Performance of nRENUM and compared framework 23
• 2.6 Conclusion 25
• 3 A Multi-Timescale Cross-Layer Approach forWireless Ad Hoc Networks 27
• 3.1 Introduction 27
• 3.2 Related Work 31
• 3.3 System Model 34
• 3.3.1 Network Model 34
• 3.3.2 Average Delay 36
• 3.3.3 Rate-Outage Probability and Effective Rate 3.6
• 3.3.4 Optimization Problem 38
• 3.4 MTSRENUM Distributed Algorithm 42
• 3.4.1 Short-timescale Iterative Subalgorithm 44
• 3.4.2 Mid-timescale Iterative Subalgorithm 46
• 3.4.3 Long-timescale Iterative Subalgorithm 48
• 3.4.4 Convergence Analysis 50
• 3.5 Simulation Results 53
• 3.5.1 Simulation Settings 54
• 3.5.2 Performance of MTSRENUM 54
• 3.6 Conclusion 60
• 4 Discussion and Future Work 62
• 4.1 Contributions 62
• 4.2 Future Work 63
• References 64