0809 Uncertainty Quantification in Particle-Based Simulations of Fluids, Polymers, and Soft Matter

  • Gerald Wang, Carnegie Mellon University

The computational study of fluids, polymers, and all manner of soft matter makes extensive use of particle-based simulations (e.g., molecular dynamics, Langevin dynamics, numerous flavors of Monte Carlo simulations, and coarse-grained variations thereof). In contradistinction to glassy and/or hard material systems, it is within the realm of imaginability that simulated dynamics of (certain) fluids, polymers, and soft matter may accurately sample a meaningful portion of the system’s phase space with tractable computational effort, leading to unbiased estimates of thermodynamic, transport, and material properties. However, a challenge common to all particle-based methods is quantifying and propagating uncertainties in these simulation measurements. In (high-level) principle, the statistical sampling procedures typically used in particle-based simulations imply that the associated uncertainties are governed by the Central Limit Theorem; in practice, faithful and efficient estimation of uncertainties in particle-based simulations is replete with challenges that require insights from a broad range of fields, including statistics, statistical physics, dynamical systems, high-performance computing, and computer science. As the applications of particle-based methods to study soft matter proliferate, and as these methods are increasingly used in tandem with (or even to drive) experimental research efforts, it is more critical than ever to nucleate conversations about uncertainty quantification amongst molecular-simulation practitioners. We propose a multi-disciplinary mini-symposium to bring researchers from several of these fields together to highlight approaches for uncertainty quantification in soft-matter systems. This mini-symposium would be a strong fit for WCCM, not only because it would attract abstracts from researchers interested in computational uncertainty evaluation and error estimation, but also because this session would be greatly beneficial for conference attendees interested in materials by design; fluid dynamics and transport phenomena; and atomistic, nano-, and micro-mechanics of materials.

© WCCM-APCOM 2022. All Rights Reserved.