Wigner function of a Schrödinger cat state. The moving computational particles (shown in amber) approximate the classical probability distribution determined by a hybrid quantum–classical wavefunction, which itself approximates the quantum state and its coupling to a classical environment.
My research lies at the interface of numerical analysis and quantum dynamics. I develop and analyse numerical methods for the time-dependent Schrödinger equation and for mixed quantum-classical systems, aiming to bridge rigorous mathematics with applied sciences such as quantum chemistry and quantum computing. Across these projects, I pair theoretical error analysis with numerical experiments to establish both the correctness and the practical performance of new methods. My work spans the following areas:
Geometric and variational methods, including hybrid wavefunctions and particle closures, with applications to nonadiabatic dynamics and physical systems such as Rashba spin-orbit coupling
Error analysis and numerical simulation of methods for the linear and nonlinear Schrödinger equation
Analysis of the approximation of wave functions via Fourier analysis, frozen and thawed Gaussian propagation for the linear Schrödinger equation
Structure-preserving time integrators for differential equations based on operator splitting
Compressed representations for high-dimensional quantum dynamics and optimization
Variational methods for near-term quantum computing
with W. Bauer, F. Gay-Balmaz, C. Tronci
SIAM Multiscale Model. Simul., 22(4):1365-1401
with C. Lasser
Numer. Math., 152, 511-551
with M. B. Soley, A. Gorodetsky, V. S. Batista
J. Chem. Theory Comput., 18, 25-36
International Meetings on Differential Equations and Their Applications
Lodz University of Technology, Online · 18 November 2026
Abstract
Operator splitting methods provide efficient approximations to evolution equations of the form $u'(t)=(A+B)u(t)$, particularly when the flows generated by the operators $A$ and $B$ can be computed separately. In applications to PDEs, however, these operators are typically unbounded: differential operators such as the Laplacian are defined only on suitable domains and generate strongly continuous semigroups rather than bounded operator exponentials. Consequently, arguments based on norm-convergent Baker–Campbell–Hausdorff series or direct Taylor expansions are generally no longer applicable, or require unnecessarily strong regularity of the initial data. In this talk, we analyze symmetric Zassenhaus splittings in both the bounded and unbounded settings. Introduced in 2014 for the time-dependent semiclassical Schrödinger equation, this family of methods has proved particularly effective in the semiclassical regime and has become an important tool for the numerical simulation of quantum dynamics. We present a rigorous error analysis of symmetric Zassenhaus splittings in the presence of unbounded operators $A$ and $B$. To overcome the limitations of formal infinite expansions, we employ an integral representation of the local error together with finite commutator expansions whose remainders are controlled explicitly. This approach identifies sufficient regularity conditions for recovering the local convergence order known from the bounded setting. It also reveals an “overshoot” phenomenon: controlling the remainder may require commutators of greater length — and hence stronger regularity — than the formal truncation order alone would suggest.