Optimization of Regenerator Placement in Optical Networks Using Deep Tensor Neural Network

Responding to the rapid growth of network traffic, the development of new backbone network technologies is essential. One of the solutions is to use Elastic Optical Networks (EONs) instead of traditional Wavelength Division Multiplexing (WDM). EON provides the means to allocate spectrum bandwidth flexibly in highly congested networks. However, due to physical limitations, many EON nodes require the use of regenerators. To optimize the cost and the efficiency of signal regeneration, we propose the use of a deep learning model. We use it to calculate the best allocation of regenerators in the network so that even with limited resources, EONs can still efficiently route traffic. After numerous experiments, our approach shows significant improvement in reducing both spectrum and regenerator blockage rates.