A Deep Learning-Based Routing and Spectrum Allocation (RSA) Strategy for Elastic Optical Networks

Abstract

With the rapid expansion of cloud services and the increasing demand for high-bandwidth applications, traditional Wavelength Division Multiplexing (WDM) networks are struggling to meet modern data transmission needs. Elastic Optical Networks (EONs) have emerged as a promising solution, offering more efficient spectrum utilization through finer resource division. In response to challenges like delay in multipath routing and the limitations of traditional RSA approaches, this paper introduces a deep learning-based strategy to optimize routing and spectrum allocation (RSA) in EONs. By incorporating a deep neural network (DNN) model, the proposed strategy dynamically adapts to real-time network conditions, reducing spectrum fragmentation and minimizing bandwidth blocking probability. Simulation results demonstrate the superior performance of this approach, marking a significant advancement in the intelligent management of next-generation optical networks.