A GA approach for traffic matrix estimation
Yi Lin Jiang, Shang Fengjun, Zou Yang, Lin-hao Li · 2009
Traffic matrix is important for many network design, engineering, and management functions. However they are often difficult to measure directly. Because networks are dynamic, analysis tools must be adaptive and computationally light weight. In order to estimate the traffic matrix for whole network, a novel calculating model is proposed based the genetic algorithm (GA). Firstly, a generalized inverse matrix is introduced to acquire the general solutions of traffic matrix equation. Secondly, in order to improve the method, an original traffic matrix is estimated according to the prior, for example, Poisson model. Lastly, genetic algorithm is proposed to estimate the traffic matrix. Through both theoretical analysis and simulating results, it is shown that the proposed algorithm achieves better performance than the existing representative methods.