Dr. Hannah Kravitz
Email: hkravitz@pdx.edu
Office: (503) 725-9228
2022 – present: Assistant Professor, Portland State University
2022: Ph.D. Applied Mathematics, University of Arizona
2014 – 2016: Data Analyst, Institute for Health Metrics and Evaluation (IHME), University of Washington
2011: B.S. Bio-Engineering The Ohio State University
I am an Assistant Professor of Mathematics in Portland State University's Fariborz Maseeh Department of Mathematics + Statistics.
My research lies at the intersection of computational mathematics, partial differential equations, network science, and public health applications. My work combines numerical analysis, PDE theory, and network-based modeling to study how structure and geometry influence dynamical behavior. Representative scholarship is provided below. For a full list of papers, please see my Google Scholar page.
Metric graphs: Building on my 2022 dissertation on numerical methods for metric graphs, I continue to develop numerical methods and mathematical theory for PDEs on these structures. My interests include numerical methods for eigenvalue problems, spectral properties of metric graphs (such as distribution of eigenvalues, localization, and steady-state behavior), and applications to epidemiology and optical science.
Mathematical epidemiology: I study how population movement along network structures (such as road networks, airline networks, and river networks) affects the geographic spread of disease. My work emphasizes PDE-based models with the goal of understanding how particular structures within networks enhance or inhibit transmission.
Coupled ODE+1D+2D systems: I develop and analyze coupled systems in which the vertices are modeled using time-dependent ODEs, the edges by one-dimensional PDEs, and the surrounding regions by two-dimensional continuum models. In this framework, each nth dimensional component acts as a boundary for the (n+1)st dimension, leading to novel numerical and analytical challenges, as well as flexible modeling capabilities for multiscale systems.
Numerical methods for irrational rotations: I recently developed a numerically stable algorithm that computes the discrepancy of an irrational rotation exactly (up to machine precision) over its entire domain). By tracking the discontinuities rather than relying on sampling, this approach allows the exact computation of the support of the distribution over a large number of term, allowing us to deduce patterns that were previously inaccessible computationally.