The field of wireless communication is moving towards the development of more efficient and scalable resource allocation methods, driven by the increasing demand for real-time intelligent services and the emergence of 6G wireless communication. Researchers are exploring the use of learning-based methods, such as learning-to-optimize techniques, to address the computational challenges posed by traditional algorithms. Graph neural networks (GNNs) are also being investigated for their potential to reduce scheduling overhead and improve network capacity. Furthermore, there is a growing interest in developing theoretically grounded solutions that can ensure satisfaction of quality-of-service (QoS) constraints and convergence to optimal solutions. Notable papers in this area include:

• A comprehensive review of learning-to-optimize model designs and feasibility enforcement techniques, which investigates the application of constrained learning-to-optimize in wireless resource allocation systems.
• A distributed link sparsification scheme employing GNNs to reduce scheduling overhead for delay-tolerant traffic while maintaining network capacity.
• A GNN-based algorithm that integrates with a Lagrangian-based primal-dual optimization method to ensure satisfaction of QoS constraints and convergence to a stationary point.
• An agentic artificial intelligence (AI)-driven double deep Q-network (DDQN) scheduling framework for licensed and unlicensed band allocation in New Radio (NR) sidelink (SL) networks.