Using Multi-Layer Perceptron and complex network metrics to estimate the performance of optical networks

Abstract

The performance assessment of a WDM network considering physical impairments is a difficult task and is frequently accomplished by using time consuming computational simulations. On the other hand, we observed that several metrics have been proposed to assess different aspects of a network structure. In this paper we propose that a set of metrics can be combined in order to obtain a fast estimation of a WDM network performance, based on a historical database of networks. The estimator was obtained by means of the most used Artificial Neural Network (ANN) architecture, called Multi-Layer Perceptron, that was trained using the classical back-propagation algorithm. According to our results, it is possible to build an estimator based on network metrics that assess WDM networks considering the trade-off between the processing time and the precision of the results. Our study also suggests that this kind of estimator can be easily adapted to other scenarios of WDM networks since Artificial Neural Networks present interesting characteristics, such as adaptation and flexibility.

Methodology to Obtain a Fast and Accurate Estimator for Blocking Probability of Optical Networks

“…A complete reference to complex networks and topological properties can be found in [7][8][9][10][11].…”

“…The C i;j value is set with the number of lightpaths that crosses the fiber link i; j. Then, CR is defined as CR c max − c min r ; (8) in which c max is the largest value in C, c min is the smallest nonzero value in C, and r is the number of nonzero positions in C. One can observe that small CR values suggest WDM networks with reasonable load balance capability.…”

“…This type of tool would allow stake-holders to assess optical mesh networks with numerous nodes considering the physical impairments in a reasonable time. In this line of work, Araújo et al [8] proposed a preliminary approach to assess network performance based on an ANN fed by network topological properties. They used an ANN with eight input parameters in order to evaluate the blocking probability.…”

“…They used an ANN with eight input parameters in order to evaluate the blocking probability. However, no formal method was deployed to choose the best set of topological properties and physical layer characteristics to be used as inputs for the ANN in [8]. As a consequence, this preliminary proposal offers a fast way to estimate BP, but presents large residual errors for some WRONs.…”

“…Furthermore, the network performance is often very sensitive to the traffic load imposed on the network. However, Araújo et al [8] did not analyze the influence of the traffic load on the quality of the proposed model. This paper aims to present a methodology to obtain a fast and accurate tool to assess the performance of optical networks based on a suitable set of topological properties combined with information of the optical physical layer.…”