Decomposition Techniques in Mathematical Programming Applied to Energy-Aware Traffic Engineering Optimization

The growing energy consumption has already become a global concern and nowadays more than
forty countries are involved in researches and programs aiming to develop mechanisms related to
energy saving. This work deals with the Energy-aware Traffic Engineering problem applied to the
backbone of an IP network in which the used routing protocol is a SPF (Shortest Path First) one.
The proposed approach considers the problem of turning off nodes (routers) and circuits, for energy
saving; and it also considers the problem of ensuring a maximum circuit utilization level, towards
to assure QoS requirements. The core idea of the approach here proposed is, rather than solving
the formulated optimization problem by means of heuristic methods, finding the best solution by a
direct processing applying Benders decomposition, which crumbles a complicated and hard to solve
problem into several smaller ones whose resolution is simpler and whose convergence is faster.