Cross-Chiral Phononic Crystal Inverse Design Based on Flexible Adaptive Marine Predator Algorithm

To address the issue of low efficiency in the design of low-frequency bandgaps in traditional phononic crystals, this paper proposes a data-driven inverse design framework. First, a novel cross-chiral phononic crystal structure is constructed, and finite element analysis is used to characterize its low-frequency bandgap characteristics and the influence patterns of key geometric parameters. We develop a high-precision multi-layer perceptron (MLP) model to establish a rapid mapping relationship between geometric parameters and bandgaps. The trained neural network is then integrated with an improved flexible adaptive marine predator algorithm (FAMPA) to achieve efficient parameter optimization. This method can precisely generate structural parameters based on the target bandgap, significantly improving design efficiency and providing a new paradigm for low-frequency vibration control.