Computational Techniques for Simulating Light Propagation in High-Energy Neutrino Telescopes

摘要

To maximize the accuracy of background simulation and event reconstruction,high-energy neutrino telescopes require detailed knowledge of light propagation over a large volume of detection medium.If light scattering and absorption lengths in the medium are of the same scale as the detector size,this problem can only be handled numerically.Any inhomogeneity of optical properties in the medium further complicates the problem,requiring large computational resources.We present a treatment based on combining ray-tracing Monte Carlo and neural network techniques which offers a reasonable compromise between solution accuracy,computer memory and CPU usage.