A Novel Genetic Algorithm for Topology Aggregation in Multidomain Optical Networks

In a optical multidomain network, each domain needs to advertise its topology information to the other domains. Topology aggregation is an effective means to reduce the amount of the advertised topology information for each domain in large scale networks. The accuracy of aggregating the topology states is improved by using the approximation curves with limited amount of advertising topology information. Polyline approximation has the best performance for all approximation methods. Based on study of the polyline approximation methods and aimed at minimizing the approximation distortion, this paper proposes a genetic algorithm to resolve the polyline approximation curve and addresses some key issues, including the coding, initial population production, fitness, crossover and mutation within such an algorithm. Simulations show that the approximation curve distortion of proposed algorithm is least comparing all other approximation curve methods.