Spectral shape optimization and asymptotic analysis of metastability for accelerated MD

ECCOMAS PhD Olympiads, Munich 2026

Noé Blassel

Mathematics for Materials Modelling group, EPFL, Lausanne, Switzerland

July 21, 2026

Molecular dynamics and the timescale problem

$$ % % % % % % % % % % % % %

% Blackboard Bold

% Calligraphic % Positive definite symmetric matrices

% Script

% Fraktur

% Other Symbols

% Alternative definition, safer % %

% General Physics/Math

% Geometry and Neighborhoods

% Hessian Related

% Harmonic Oscillators

% Simplified/Indexed Notation

% Other Specific Notation

$$

\[ \definecolor{gris_C}{RGB}{96,96,96} \definecolor{blanc_C}{RGB}{255,255,255} \definecolor{pistache_C}{RGB}{176,204,78} \definecolor{pistache2_C}{RGB}{150,171,91} \definecolor{jaune_C}{RGB}{253,235,125} \definecolor{jaune2_C}{RGB}{247,208,92} \definecolor{orange_C}{RGB}{244,157,84} \definecolor{orange2_C}{RGB}{239,119,87} \definecolor{bleu_C}{RGB}{126,151,206} \definecolor{bleu2_C}{RGB}{90,113,180} \definecolor{vert_C}{RGB}{96,180,103} \definecolor{vert2_C}{RGB}{68,141,96} \definecolor{pistache_light_C}{RGB}{235,241,212} \definecolor{jaune_light_C}{RGB}{254,250,224} \definecolor{orange_light_C}{RGB}{251,230,214} \definecolor{bleu_light_C}{RGB}{222,229,241} \definecolor{vert_light_C}{RGB}{216,236,218} \]

Molecular Dynamics (MD) simulations act as a numerical microscope when experiments are unfeasible/unaffordable

Applications include drug design and materials discovery

  • Motion of atomic nuclei modelled with a stochastic process \(X\) on space \(\mathcal{X}\) of atomic configurations

  • Stationary distribution is the Gibbs measure \(\mu(\mathrm{d}x) = Z_\beta^{-1}\mathrm{e}^{-\beta V(x)}\,\mathrm{d}x\) for the potential energy \(V:\mathcal{X}\to\mathbb{R}\) and inverse temperature \(\beta = 1/(k_{\mathrm{B}}T)\)

  • Thermodynamic averages require samples under the Gibbs measure

  • Dynamical quantities (reaction rates, transport coefficients, autocorrelations…) require long unbiased trajectories, and are much harder to estimate

  • Main obstruction is metastability

  • Simulation step \(\approx 1\,\mathrm{fs}\), interesting kinetic events \(\gtrsim 1\,\mu\mathrm{s}\implies\) timescale problem