Algorithms in the deployment of Optical Transport Networks

We propose a multi-thread multi-start local search strategy to optimize the cost of the next generation of Optical Transport Networks (OTNs) in the presence of heterogeneous client demands and considering a wide range of engineering solutions and constraints whose combination is hard to address, such as flexible grid frequency assignment, single and multi-hop grooming, regenerator placement and inverse-multiplexing. The multi-thread approach enables the computational resources of current multi-core platforms to be fully utilized. The computational results show that the algorithm´s efficiency improves when the information of the best solution found by any thread is exchanged among all threads, making this method a good approach to deal with highly complex optimization problems.