Description
Accurate simulation of molecular vibrational quantum dynamics towards rigorously solving the multidimensional Schrödinger equation with direct grid-based methods is severely limited by the required storage space and computational effort, which scale exponentially with the number of coupled degrees of freedom. In this talk, I will show how to use tensor networks to solve partial differential equations, focusing on how to adapt these techniques for simulating complex vibrational dynamics. By treating the function over the grid as a tensor network, this approach provides data compression. We will first benchmark this methodology by calculating the infrared spectrum of water and formaldehyde under the second-order vibrational perturbation approximation. I will then show how to extend the method to larger molecules while including the exact potential energy surface.