On the efficient column generation-based optimization of anycast traffic in survivable elastic optical networks

In this paper, we focus on efficient column generation (CG)-based method for routing and spectrum allocation (RSA) of anycast traffic in survivable elastic optical networks (EONs). The studied anycast traffic is modelled as a two-directional transmission (from the client node to a data center and from a data center to the client) while the network is protected against a single link failure by a dedicated path protection (DPP). The paper main focus is put on proposal and comparison of different policies regarding column adding process for anycast demands. Overall, four CG versions are proposed. The simulation results show that the versions significantly differ in the performance especially in terms of the processing time. What is more, the results prove high efficiency of the proposed CG approach that outperforms reference methods and achieves solutions very close to optimal ones (average gap to optimal result was at most 2%).