By Sousa, P.; Cortez, P.; Rio, M.; Rocha, M.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Nowadays, network planning and management tasks can be of high complexity, given the numerous inputs that should be consid- ered to effectively achieve an adequate configuration of the underlying network. This paper presents an optimization framework that helps net- work 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 infras- tructure. Since the envisaged task is a NP-hard problem, the framework resorts to Evolutionary Computation as the optimization engine. The fo- cus is given to the use of multi-objective optimization approaches given the flexibility they provide to network administrators in selecting the ad- equate solutions in a given context. Resorting to the proposed optimiza- tion 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 man- agement tasks.