National Technical University of Athens (NTUA)
Διαστρωματική σχεδίαση και βελτιστοποίηση ασύρματων πολυβηματικών δικτύων
Abstract
dc:descriptionThis dissertation focuses on the improvement of the communication provided by wireless multihop networks following a holistic approach with respect to the protocol stack. Specifically, the focus of this dissertation lies in the development of two cross-layer frameworks that contribute either combinatorially, or each one independently in the improvement of Quality of Service (QoS) metrics. The first framework is denoted as Inverse Topology Control Framework (iTC Framework), as being based on topology control, and functions on Physical and Network layers while leveraging information and features from the social layer. The iTC Framework enhances the topology of the wireless multihop network with features from social graphs/Social Network Analysis (SNA), following the paradigm of the efficient and fast communication over social networks. We propose the probabilistic operation of the iTC Framework, we develop an optimization methodology of its parameters and we describe the prospects of its distributed implementation. The analytical study of the iTC Framework is based on Continuum Theory. Furthermore, we present the implementation of the iTC Framework in special cases and applications such as trust and traffic management which show its effectiveness in improving several QoS metrics e.g., delay, throughput, trust and cost of packet transfer.The second framework is denoted as TNUM/Backpressure Framework as it is founded on the Theory of Network Utility Maximization (TNUM) and functions on MAC, Network and Transport layers. The optimal solution of a NUM problem is computed based on the Backpressure (BP) cross-layer algorithm for scheduling and routing along with an appropriately defined congestion controller for the determination of the optimal arrival rates' values at the sources. The BP algorithm is optimal with respect to the capacity region (throughput optimal, i.e., it leads to stable network queues for all arrival rates lying inside the capacity region) but leads to high delay values and does not provide the possibility of considering other Quality of Service metrics. In this work, the TNUM/Backpressure Framework introduces appropriate weights in the network graph and the greedy embedding of the network graph in hyperbolic space. The BP algorithm is adapted for taking into consideration the weights' values, denoted as weighted BP, and via introducing suitable routing constraints for static and dynamic networks. More specifically, with respect to delay reduction, we apply greedy routing constraints in hyperbolic space. We examine the stability of the proposed weighted BP algorithm and the convergence and stability of the proposed algorithm for solving the NUM problem, which is based on the weighted BP, using primal-dual decomposition and Lyapunov drift techniques. The simulation results evaluate the possibility of delay reductions and other QoS metrics' improvements (e.g., cost, trust, etc.). Furthermore, we examine via simulations the effect of weights on QoS, on the convergence and stability of the NUM problem's algorithmic solution, while also studying sophisticated choices for the weight values for each QoS metric. In addition, the greedy routing in hyperbolic space is improved by applying coalitional game theory aiming to further boost the performance of the proposed algorithms by cooperatively deciding for the greedy embedding in hyperbolic space. In the sequel, the NUM problem is formulated and solved so that the trade-off between the sources' arrival rates (long-term throughput) and delay becomes users' controlled. Then, social layer's information is used for the utility function design targeting at a more efficient resources' distribution among users according to their needs.Finally, the two proposed frameworks are applied together so as to study, via examples and simulations, the possible combinatorial improvement in QoS. At the end of the dissertation we summarize this research's basic conclusions and we pinpoint and discuss the current research along with possible future steps.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2015
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Stai, Eleni
- Στάη, Ελένη
Subjects
dc:subject × 20- Ασύρματα πολυβηματικά δίκτυα
- Διαστρωματική σχεδίαση και βελτιστοποίηση
- Κοινωνικά δίκτυα
- Ποιότητα υπηρεσίας
- Έλεγχος τοπολογίας
- Θεωρία μεγιστοποίησης χρησιμότητας δικτύου
- Wireless multihop networks
- Cross-layer design and optimization
- Social networks
- Quality of service
- Topology control
- Network utility maximization
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Natural Sciences
- Computer and Information Sciences
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
dc:rights- Statement dc:rights
-
- BY_NC_ND
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/38647