Intelligent Maintenance Scheduling for 5G Networks with ECA-DenseNet-DQN

This paper presents an intelligent maintenance scheduling model for 5G networks using a deep reinforcement learning approach, specifically ECA - DenseN et- DQ N. The proposed model optimizes the allocation of maintenance personnel by considering both travel and task completion times, aiming to minimize total task duration. Unlike traditional scheduling methods that often disregard skill levels and workloads of personnel, the model integrates Efficient Channel Attention (ECA) mechanisms into DenseNet to enhance feature extraction capabilities, thereby improving scheduling decisions. Simulation experiments demonstrate the model's superior performance in reducing task completion time and achieving faster convergence compared to existing algorithms.