A utility-aware scheduling model for online learning tasks based on dynamic psychological cognitive load perception

In online learning scenarios, the dynamic fluctuations of students' cognitive load directly impact learning outcomes.Existing scheduling models, lacking real-time perception of cognitive load states, often result in mismatches between task assignments and learner capabilities.To address this, this paper first processes student contextual features using feature selectors and self-attention mechanisms, then predicts response performance based on cognitive load state.Subsequently, an online learning task utility scheduling model is constructed based on cognitive load diagnosis results.A multidimensional utility function for online learning tasks is designed, establishing a scheduling objective function that maximises this utility.Finally, an improved particle swarm optimisation algorithm solves the objective function to derive the optimal online learning task utility scheduling strategy.Experimental results demonstrate that the proposed method achieves scheduling times of 3.3 ms and a success rate of 98.3%, outperforming baseline methods with significantly higher scheduling efficiency.