• Optimization
• EPC
• LTE
• Routing
• Network

The 3GPP Evolved Packet System (EPS) is the new system for mobile communications. Such system is composed at a radio access, called Long Term Evolution (LTE), and an IP backbone architecture called Evolved Packet Core (EPC). At high-level view, flows are rep- resented by bearers. A bearer is a virtual path that connects a user equipment to the egress gateway of the EPS, that is the Packet Data Network Gateway (PDN-GW). The topic investigated within this the- sis is how to route flows in the EPC so as to meet the QoS guarantees included in the bearer while optimizing network resources. In the EPC, bearer can be pre-empted when network resources are scarce. Therefore, route computation should also be judged by the number of pre-emptions that that it forces in the network. Three different routing optimization models with different behaviours were developed within this thesis which incorporate pre-emption in different ways. We have tested these models in realistic networks as the blocking proba- bility, pre-emption probability and time overhead. Our results show that the algorithms exhibits similar behaviours as the blocking and pre-emption probability. Moreover, blocking and pre-emption seems to be contrasting goals one is reduced at the expenses of the other. As for time overhead, computations on off-the-shelf hardware show that an optimization approach is compatible with a highly-dynamic large-scale environment.