National Technical University of Athens (NTUA)
Αποδοτικοί αλγόριθμοι σχεδίασης δικτύων επικοινωνιών με έμφαση στην ποιότητα υπηρεσίας
Abstract
dc:descriptionNetwork planning is the use of scientific methods for optimizing the investment and for dimensioning network equipment in a unified way’ for a whole system, in order to meet realistic business objectives previously defined by the highest level authorities, according to International Telecommunication Union. The challenge of optimal planning is significantly increased in Next Generation Networks when traffic is actively segmented into different traffic classes. Utilizing Traffic Engineering along with effective optimization algorithms is the answer for network design in modem communication networks. Due to the NP-hard nature of network design problems, meta-heuristics provide a viable optimization alternative option. The dissertation focuses on the topological design and dimensioning of Layer-3 communication networks with Quality of Service considerations, utilizing particle swarm optimization. Initially, an instance of a real telecommunication best-effort network design problem is considered as a single objective optimization problem using two variations of Particle Swarm Optimization. Their effectiveness on solving the aforementioned problem is checked against the application of a genetic algorithm. The problem under consideration is expanded with Quality of Service constraints (i.e. delay). It is formulated as a bi-criteria optimization problem, where multi objective Particle Swarm Optimization is utilized as the optimization technique. Finally, the design of multiclass - tree-like distribution networks is considered. The target is to minimize network deployment cost and end-to-end delay in the network per traffic class, subject to pre-defined bounds. This is presented as two separate problems; i. a scheduling design and verification problem and ii. a topological design problem extended by including the scheduling configuration parameters in the optimization process. Whereas in literature mostly Priority Queuing systems have been considered in design efforts as such, the current work extends the problem by adopting class based weighted fair queuing with low latency queue as the scheduling policy. The choice of the scheduler for a multi-service communication network is validated by simulation of indicative distribution/aggregation broadband architecture. The topological design problem is formulated as a constrained multi-objective problem and multi objective Particle Swarm Optimization is applied. Its efficiency is evaluated against Non-dominated Sorting Genetic Algorithm II.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2009
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Παπαγιάννη, Χρυσούλα
- Papagianni, Chrysoula
Subjects
dc:subject × 24- Βελτιστοποίηση σμήνους σωματιδίων έλξης απώθησης
- Πολυστοχική βελτιστοποίηση
- Βελτιστοποίηση μονού στόχου
- Πρόβλημα τοπολογίας - διαστασιολόγησης - ανάθεσης ροών
- Δίκαιη αναμονή με βάρη - ουρά χαμηλής καθυστέρησης
- Προσομοιώσεις
- Πακέτο υπηρεσιών IP
- Διαφοροποιημένες υπηρεσίες
- Particle swarm optimization
- Multiobjective optimization
- Single objective optimization
- Topology capacity and flow allocation
- Weighted fair queving - low latency queue
- Simulation
- Triple play service
- Diffserv
- Quality of service
- Hybridization
- Attractive- repulsive PSO
- NetKeys Encoding
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/18613