Unit:
Τομέας Επικοινωνιών και Επεξεργασίας ΣήματοςLibrary of the School of Science
Author:
Μανιαδάκης Δημήτριος
Dissertation committee:
Δημήτριος Βαρουτάς Επίκ. Καθηγητής (επιβλέπων), Λάζαρος Μεράκος Καθηγητής, Θωμάς Σφηκόπουλος Καθηγητής
Original Title:
Complex patterns in telecommunications network planning
Translated title:
Πολύπλοκα πρότυπα στο σχεδιασμό τηλεπικοινωνιακών δικτύων
Summary:
This PhD thesis discusses a series of related issues raised towards the
efficient telecommunications network planning, taking into account the
existence of complexity. Three aspects of the associated complexity are
examined, focusing on both the telecommunications networks per se, and the
underlying street networks: (i) the observation of complexity, (ii) its effects
and (iii) its utilization. More specifically, the empirical findings which
demonstrate the existence of complex connection patterns are initially
described. By collecting novel datasets of network data and graph-theoretically
analyzing them, specific patterns of complex connectivity are identified. Next,
the impact of complexity in the network efficiency is examined. The comparison
of both synthetic and real-world topologies in terms of availability,
congestion and cost, shows significant differences and indicates the
criticality of specific topological properties. Especially, concerning the
traffic congestion, the inclusion of topologies based on the gravity model
reveals their superiority. Subsequently, a novel methodology is proposed, using
synthetic complex graphs for the preliminary design of fiber-to-the-x
telecommunications access networks. In particular, the utilization of planar
proximity Gabriel graphs to represent the underlying layer in an urban
environment is proved to be superior to the conventional geometric dimensioning
models.
Keywords:
Complex networks, Statistical network analysis, Optical access networks, Street networks, Network planning
Number of index pages:
29
Number of references:
120
File:
File access is restricted.
document.pdf
7 MB
File access is restricted.