Naas, Nabil A2013-11-082013-11-0820082008Source: Dissertation Abstracts International, Volume: 70-08, Section: B, page: 5060.http://hdl.handle.net/10393/29741http://dx.doi.org/10.20381/ruor-19885The explosive growth of traffic in Wavelength Division Multiplexing (WDM)-based transport networks has necessitated an increase in the number of wavelengths per fiber. However, the continuation of this trend will definitely lead to both an explosion in the size of Optical Crossconnect (OXCs)/Reconfigurable Optical Add/Drop Multiplexers (ROADMs) and an increased difficulty in routing and managing wavelengths. With the introduction of Generalized Multiprotocol Label Switching (GMPLS), the multi-granular switching concept becomes essential to alleviating these serious scalability problems. However, in spite of the fact that this new switching concept brings many benefits, it also results in new network design and planning challenges. In this thesis, a new problem in planning the GMPLS-based (or multi-granular) transport network, called the Routing and Multi-Granular Paths Assignment (RMGPA) problem, is addressed. The objective of the RMGPA problem is to minimize the total weighted port count in the transport network. The novelty of this problem lies in the incorporation of the following for the first time: (1) considering the whole traffic hierarchy defined in GMPLS; (2) optimizing multi-granularity traffic demands; (3) allowing the bifurcation of multi-granularity demands; (4) (de)grouping traffic flows at all granularity levels; (5) allowing wavelength, waveband, and fiber conversions; (6) imposing the optical-reach constraint on the length of all-optical paths; and (7) customizing the optical reach of all-optical paths. In addition, in order to realize the functionalities required by these newly addressed issues, a new GMPLS-based node architecture is proposed. Different versions of the RMGPA problem are formulated as Mixed Integer Linear Programming (MILP) and Non-linear Programming (NLP) models. Due to the computational complexity of the planning problem, only the MILP models of small-sized problems are solved. To satisfy the planning speed requirement, various efficient heuristics that are capable of solving large-sized problems in a reasonable amount of time are also proposed. In addition, the solutions to the MILP models are used to validate the proposed heuristics. Moreover, in order to achieve the best possible planning solution to different versions of the RMGPA problem in both fiber-scarce and plentiful situations, a comprehensive evaluation of different heuristic variations through illustrative examples and simulations is conducted. Furthermore, the results provide valuable insights into many issues that can contribute to further research and development in this area.355 p.enEngineering, Electronics and Electrical.Thesis