Research on AI-Driven Optimization Algorithm for Educational Resource Scheduling in Smart Campus

We present a deep reinforcement learning-based scheduling algorithm framework to optimize educational resource allocation in smart campuses. The framework starts with a multi-dimensional resource feature model that considers spatiotemporal constraints and user preferences. The core algorithm uses a deep reinforcement learning network with dual attention mechanisms, which helps process heterogeneous resource features. We developed a scheduling system that can scale and adjust in real time. Tests reveal that our approach achieves better performance than conventional methods. The system maintains resource utilization between 80% and 90% in high-load conditions, with response times ranging from 45ms to 92ms. The framework also shows flexibility across various scenarios, helping smart campuses manage their educational resources effectively.