Date of Award


Publication Type

Master Thesis

Degree Name



Computer Science

First Advisor

Jaekel, A.


Computer Science.




Internet development generates new bandwidth requirement every day. Optical networks employing WDM (wavelength division multiplexing) technology can provide high capacity, low error rate and low delay. They are considered to be future backbone networks. Since WDM networks usually operate in a high speed, network failure (such as fiber cut), even for a short term, can cause huge data lost. So design robust WDM network to survive faults is a crucial issue in WDM networks. This thesis introduces a new and efficient MILP (Mixed Integer Linear Programming) formulation to solve dynamic lightpath allocation problem in survivable WDM networks, using both shared and dedicated path protection. The formulation defines multiple levels of service to further improve resource utilization. Dijkstra's shortest path algorithm is used to pre-compute up to 3 alternative routes between any node pair, so as to limit the lightpath routing problem within up to 3 routes instead of whole network-wide. This way can shorten the solution time of MILP formulation; make it acceptable for practical size network. Extensive experiments carried out on a number of networks show this new MILP formulation can improve performance and is feasible for real-life network. Source: Masters Abstracts International, Volume: 43-01, page: 0249. Adviser: Arunita Jaekel. Thesis (M.Sc.)--University of Windsor (Canada), 2004.