Multi-Objective Integer Linear Programming Model for Path Optimization in Wavelength Division Multiplexing Networks

Abstract

Optical networks with Wavelength Division Multiplexing (WDM) has been the solution for the need of increasing bandwidth demand. In this kind of networks the fiber link is divided into a number of channels. In each channel a light wave of a particular wavelength can be transmitted. So, in a single fiber more than one light waves of different wavelengths can be transmitted simultaneously with the use of multiplexers and demultiplexers. In WDM optical networks there are two problems. First one is providing a path for a source to destination pair, and second one providing a wavelength to the path selected. The former is called routing problem and the latter is called wavelength assignment problem. Combining both the problems, it is called Routing and Wavelength Assignment (RWA) Problem. The RWA problem belongs to NP class, i.e. it can not be solved in polynomial time. So, different heuristic approaches are used to find a (sub)optimal solution for the problem. An ILP model may be used to solve the RWA problem considering various parameters of the optical network like congestion, total route length, number of amplifiers used etc. In this project an ILP is designed for the RWA problem and is solved using genetic algorithm to find an optimal solution. The simulation is carried out on Advanced Research Project Agency NETwork (ARPANET) and National Science Foundation Network(NSFNET).