Provisioning and Recovery in Flexible Optical Networks using Ant Colony Optimization

Artificial intelligence techniques may play a significant role in the control of elastic optical networks by offering improved performance over traditional dynamic provisioning and recovery methods. In this work, we propose a fully distributed routing algorithm based on the Ant Colony Optimization (ACO) meta-heuristics associated with crankback re-routing extensions. Simulations have demonstrated that the proposed ACO-based algorithm outperforms the fixed-alternate approach. They also demonstrate that the proposed approach achieves similar levels of bandwidth blocking probability and restorability compared to an adaptive routing algorithm that relies on the OSPF-TE routing protocol, with much lower levels of control overhead, comparable levels of setup time, and without a significant increase in the restoration time. We also evaluate the trade-offs in using a trunk reservation policy for the proposed approach.