| dc.contributor.author | Goodridge, Wayne Sylvester. | en_US |
| dc.date.accessioned | 2014-10-21T12:35:36Z | |
| dc.date.available | 2005 | |
| dc.date.issued | 2005 | en_US |
| dc.identifier.other | AAINR13050 | en_US |
| dc.identifier.uri | http://hdl.handle.net/10222/54764 | |
| dc.description | The integration of traffic from diverse applications in core networks is placing higher demands on service quality of networks. Traffic Engineering provides a means of controlling traffic through a network by intelligent traffic mappings based on multiple traffic constraints and network carriers optimization goals. However, the problem of finding paths that satisfy the QoS requirements of traffic flows is an NP-complete problem. | en_US |
| dc.description | The major contribution of this thesis is the proposal of a multiple constraint optimization algorithm called routing decision system (RDS) that uses the concept of preference functions to address the problem of finding paths in core networks that satisfy traffic-oriented performance problems while simultaneously satisfying resource-oriented performance problems. The RDS algorithm is used in conjunction with an exact algorithm called Constraint Path (CP) algorithm or a heuristic algorithm called Constraint Path Heuristic (CP-H) algorithm both of which are novel approaches introduced in this thesis to find a set of constraint paths between source and destination nodes in a network. | en_US |
| dc.description | A comprehensive Mathematical analysis of the CP and RDS algorithms is given and it is shown that the CP algorithm is an exact algorithm, and therefore guaranteed to find a path that meets the specified constraints, provided that such a path exists. The RDS algorithm is shown to always find a pareto optimal path provided such a path exists in the set of constraint paths. This result is significant since the primary purpose of the RDS is to find optimal paths based on user optimization goals. | en_US |
| dc.description | In this thesis extensive simulations are performed to compare both the exact and heuristic versions of the CP algorithm used in conjunction with the RDS algorithm with other state of the art QoS algorithms. Results show that the algorithm has a 100 percent success rate in finding paths satisfying multiple user constraints. In addition, it is found that although exact algorithms tend to have high execution times that the CP/RDS algorithm has better running times than other exact algorithms. On the other hand, the success rate of the CP-H/RDS is shown to be significantly better than other heuristic based algorithms under strict constraints. In addition, it is shown that the associated execution time of the CP-H/RDS algorithm is slightly higher than other heuristic based algorithms but good enough for use in an online TE application. (Abstract shortened by UMI.) | en_US |
| dc.description | Thesis (Ph.D.)--Dalhousie University (Canada), 2005. | en_US |
| dc.language | eng | en_US |
| dc.publisher | Dalhousie University | en_US |
| dc.publisher | | en_US |
| dc.subject | Mathematics. | en_US |
| dc.subject | Computer Science. | en_US |
| dc.title | Multiple optimization of network carrier and traffic flow goals using preference function modeling. | en_US |
| dc.type | text | en_US |
| dc.contributor.degree | Ph.D. | en_US |
