Deep Learning-Assisted Design of Microwave Absorbers with Bidirectional Impedance Gradient Matching

To achieve efficient design of bidirectional impedance-gradient microwave absorber with multi-dimensional parameters, this paper proposes a novel design method based on a dual-channel deep learning (DCDL) network. The proposed DCDL can generate dual channels to optimize physical dimensions and material parameters, and achieve designable absorption rates. The effectiveness of DCDL is validated through two examples: forward prediction and inverse design of absorbers. The proposed method provides a robust foundation for intelligent optimization of microwave absorbers with multi-dimensional parameters.