Abstract

Part 2: Performance and Simulation Analysis; International audience; Nowadays, network planning and management tasks can be of high complexity, given the numerous inputs that should be considered to effectively achieve an adequate configuration of the underlying network. This paper presents an optimization framework that helps network administrators in setting the optimal routing weights of link state protocols according to the required traffic demands, contributing in this way to improve the service levels quality provided by the network infrastructure. Since the envisaged task is a NP-hard problem, the framework resorts to Evolutionary Computation as the optimization engine. The focus is given to the use of multi-objective optimization approaches given the flexibility they provide to network administrators in selecting the adequate solutions in a given context. Resorting to the proposed optimization framework the administrator is able to automatically obtain highly optimized routing configurations adequate to support the requirements imposed by their customers. In this way, this novel approach effectively contributes to enhance and automate crucial network planning and management tasks. Document type: Part of book or chapter of book