The recent developments in the field of software-defined networking (SDN) and elastic optical networks (EONs) have shown a significant shift towards leveraging advanced machine learning and reinforcement learning techniques to optimize network performance and resource utilization. Researchers are increasingly focusing on integrating these intelligent algorithms with traditional routing and spectrum assignment methods to enhance efficiency, particularly under dynamic traffic conditions. The use of deep reinforcement learning combined with control barrier functions is emerging as a promising approach to ensure safe operations while maintaining high performance in load balancing scenarios. Additionally, there is a growing emphasis on developing robust simulation tools that not only support traditional methods but also facilitate the integration of cutting-edge machine learning techniques, thereby providing comprehensive platforms for future research. The field is also witnessing a critical evaluation of measurement instability in virtualized environments, with methodologies being developed to ensure accurate performance assessments of software routers.