A Genetic Algorithm for Optimizing Throughput in Non-broadcast WDM Optical Networks
We apply a genetic algorithm approach to a virtual topology design of a Wide-Area WDM Optical Network. Based on a given physical topology (existing optical cables) a virtual topology consisting of a set of optical lightpaths is constructed. The objective is to minimize the maximal throughput, which implies balancing the traffic loads on network links and accommodating on-growing traffic requirements in a timely fashion. A heuristic algorithm based on genetic strategy is developed and applied to an instance of the traffic through the T1 NSFNET. A nonlinear mixed-integer programming formulation is linearized by fixing up-front the variables representing wavelengths used by optical paths. Integer matrices representing the allocation of wavelengths are iterated through the generations using an appropriate crossover function. We start with a large number of wavelengths to assure the intermediate feasibility. The final solution is obtained by means of a wavelength reduction heuristic procedure. A comparison with results from the literature is performed based on several solution quality measures. In particular, if compared with the result for a 14-dimensional NSFNET data instance from the literature, we have improved the scale factor by 74%.
KeywordsGenetic Algorithm Optical Network Physical Link Virtual Link Wavelength Assignment
Unable to display preview. Download preview PDF.
- 1.Goldberg, D.E. (1989) Genetic algorithms in search, optimization, and machine learning. Addison-Wesley, New YorkGoogle Scholar
- 4.Mukherjee, B. (1997) Optical communication networks. McGraw-Hill, New YorkGoogle Scholar
- 5.Reuters (1999) Study: More than 100 million Americans use net. done by Strategis Group Google Scholar
- 7.Skorin-Kapov, J., Labourdette J.-F. (1997) On throughput optimization in optical networks with regular and arbitrary topologies. Journal of Computing and Information Technology 5/4, 219–230Google Scholar