An Optimization Method for Wideband Bandpass Filters Design Using Autoencoder and Grey Wolf Optimizer

In this paper, an optimization method for structural parameters of wideband bandpass filters (BPFs) design based on autoencoder (AE) and grey wolf optimizer (GWO) is proposed. Once the AE is trained, the encoder of the AE can quickly predict the structural parameters from the S-parameters, the decoder of the AE can replace the EM simulation to obtain the S-parameters from the structural parameters. In order to quickly obtain the optimal structure parameters, the GWO and the decoder are used for optimization. A BPF with a pair of coupled-line and two short-circuited stub is proposed to prove the effectiveness of the proposed method. The results demonstrate that the proposed method can quickly and accurately obtain the optimal structural parameters of the BPF.