Elastic Spherical Cloak Driven by Bayesian Optimization

<正>The introduction of machine learning algorithms has revolutionized the design of invisible devices, particularly in the intricate domain of elasto-dynamics. In this study, we proposed a core-shell configuration to realize elastic sphere cloaks driven by a Bayesian optimization algorithm to pinpoint the optimal configuration with high precision. Numerical simulations in solid and aqueous environments were performed to validate the cloaking efficacy of our design and the parameters identified by the algorithm. The results closely agreed with the theoretical predictions, underscoring the robustness of the proposed method. This approach provides new insights into the design of elastic wave invisibility devices and has potential applications in underwater communication and sonar detection.