A Multi-Objective ILP Formulation for RWA Problem in WDM Networks

Abstract

All-optical networks employing Wavelength Division Multiplexing (WDM) technique will be the backbone of next generation Internet. In WDM optical networks, each fiber link is logically divided into multiple non-interfering, circuit-switched communication channels known as avelength channels and are identified by the length of the wave.Routing and Wavelength Assignment (RWA) problem is a classical problem in WDM networks. It is further divided into two subproblems: (i) Routing, and (ii)Wavelength Assignment. Routing subproblem finds a route fromsource to destination.Wavelength Assignment subproblem assigns a wavelength to the route established byrouting subproblem. The RWA problem is combinatorial by its nature and belongs to a class of dicult combinatorial optimization problems. The optimal solution to the RWA problem is found to be NP-complete and thus suited to heuristic approaches. RWAproblem is reported in the current literature as an integer linear programming problem (ILP) that typically optimizes a single objective, either minimizes the number
amplifiers, the network load or maximizes the number of connections while satisfying power constraints. In this work, we formulated the RWA problem as a multi objective
ILP problem. Our primary concern is to establish a loop free lightpath that is immune to signal distortion and crosstalk. An attempt is made to obtain a feasible solution
using genetic algorithm (GA). The parameters considered for optimization are congestion among the individual lightpath requests, connection set up time, the number of
intermediate hops traversed and the number of fibers used to honor the established connection requests. We onsidered ARPANET (Advanced Research Project Agency NETwork) and NSFNET (National Science Foundation NETwork) for our simulation.