The dense deployment of heterogeneous communication networks significantly improves spectrum utilization and network capacity but simultaneously introduces complex co-channel and cross-tier interference. To address the challenges of multi-source interference, dynamic network environments, and large-scale coordination, this study develops a collaborative framework integrating intelligent interference suppression and dynamic network optimization. A deep reinforcement learning-based interference coordination algorithm is first designed to adaptively adjust transmission power and spectrum resource allocation according to channel conditions and traffic load, thereby improving spectrum efficiency and reducing inter-layer interference. Subsequently, a federated learning-based crossdomain optimization strategy is proposed to achieve collaborative resource scheduling and load balancing without sharing raw user data. To validate the effectiveness of the proposed framework, simulation experiments are conducted under urban hotspot, high-speed railway, and industrial deployment scenarios. Results demonstrate significant improvements in interference mitigation capability, network robustness, and resource utilization efficiency. The proposed method provides technical support for future wireless communication systems and contributes to the development of electromagnetic wave propagation management, intelligent spectrum allocation, and next-generation heterogeneous networks.
Y. Li (Thu,) studied this question.