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.