Machine Learning-Based Modeling of Muon Beam Ionization Cooling

Surrogate modeling can lead to significant improvements of beam dynamics simulations in terms of computational time and resources. Application of supervised machine learning, using collected simulation data allows to build surrogate models which can estimate beam parameters evolution based on the provided cooling channel design. The created models help to understand the correlations between different lattice components and the importance of specific beam properties for the cooling performance. We present the application of surrogate modeling to enhance final muon cooling design studies, demonstrating the potential of such approach to be integrated into the design and optimization of other components of future colliders.