Rapid Implicit Diagonalization of Variable-Coefficient Differential Operators Using the Uncertainty Principle

<p>We propose to create a new numerical method for a class of time-dependent PDEs (second-order, one space dimension, Dirichlet boundary conditions) that can be used to obtain more accurate and reliable solutions than traditional methods. Previously, it was shown that conventional time-stepping methods could be avoided for time-dependent mathematical models featuring a finite number of homogeneous materials, thus assuming general piecewise constant coefficients. This proposed method will avoid the modeling shortcuts that are traditionally taken, and it will generalize the piecewise constant case of energy diffusion and wave propagation to work for an infinite number of smaller pieces, or a smoothly varying coefficient. We hypothesize that by treating a smoothly varying function as a piecewise constant function with infinitely many pieces, this potential method can be realized. Through the Uncertainty Principle, we will formulate highly accurate estimates of eigenvalues, and then through the QR algorithm, we will use these estimates to formulate highly accurate eigenvalues and eigenfunctions. Ultimately, we will produce a more efficient solution method that avoids traditional time-stepping.</p>

Name thesis:degree_name

Master of Science (MS)

Level thesis:degree_level

Masters Thesis

Year dc:date.available

2020

Author dc:creator

• Walker, Carley

Contributors dc:contributor

• Dr. James Lambers
• Dr. Haiyan Tian
• Dr. Huiqing Zhu

• partial
• differential
• equation
• smoothly-varying
• second-order
• one-dimensional
• Partial Differential Equations

Repository record dc:identifier

https://aquila.usm.edu/masters_theses/778

OAI identifier oai:identifier

oai:aquila.usm.edu:masters_theses-1830