| We are successful in building a code that models many particle dynamic quantum systems by combining a semiclassical approximation of Feynman path integrals with parallel computing techniques (particle-in-cell) and numerical methods developed for simulating plasmas, establishing this approach as a viable technique for multiparticle time-dependent quantum mechanics. Run on high-performance parallel computers, this code applies semiclassical methods to simulate the time evolution of wavefunctions of many particles. We describe the analytical derivation and computational implementation of these techniques in detail. We present a study to thoroughly demonstrate the code's fidelity to quantum mechanics, resulting in innovative visualization and analysis techniques. We introduce and exhibit a method to address fermion particle statistics. We present studies of two quantum-mechanical problems: a two-electron, one-dimensional atom, resulting in high-quality extractions of one- and two-electron eigenstates, and electrostatic quasi-modes due to quantum effects in a hot electron plasma, relevant for predictions about stellar evolution. We supply discussions of alternative derivations, alternative implementations of the derivations, and an exploration of their consequences. Source code is shown throughout this dissertation. Finally, we present an extensive discussion of applications and extrapolations of this work, with suggestions for future direction. |
