A methodology for traffic shaping multiobjective optimization for next generation network

In recent years, computer networks have been characterized by heterogeneous traffic and dynamic management of different kinds of services. The network requirements have increased within time and, since bandwidth is limited, it becomes necessary to employ optimization procedures in order to make the network able to operate in its full capacity. Traffic shaping mechanisms implement Quality of Service (QoS) concepts to ensure acceptable service levels. This paper shows preliminary results of an empirical approach for traffic shaping optimization. It is proposed a methodology based on throughput and packet loss optimization through multiobjective genetic algorithms.