Structure of fermionic wave functions and its influence on accuracy of fixed-node diffusion Monte Carlo methodSupervisor: Ing. Matúš Dubecký, PhD.
This page shows details on the project. The primary projects are displayed together with a list of sub-projects.
|Project description:||In recent years, single-determinant fixed-node diffusion Monte Carlo (FNDMC) reached high-standard accuracy in a number of diverse systems (where mean-field methods like DFT do not suffice) ranging from weakly bound noncovalent complexes to strongly correlated systems like solid transition-metal oxides at high pressures. Thanks to its favourable CPU cost scaling, parallelism, and direct access to periodicity, FNDMC gains popularity as an unprecedented benchmark tool for large realistic complex many-electron systems. Recent results however suggest, that the expected accuracy is not always accessible, sometimes the results are overvalued, or they depend on the parameters that have been, to date, ignored. The reason being incomplete understanding of FN approximation (FNA) and its interplay with other possible biases. Our goal is identification and development of deep conceptual understanding of the key FNDMC error sources. We plan to uncover the currently unknown links between generic nodal (position-space) properties (e.g., topology) of fermionic wave functions, and, their connection to the structure of many-determinant expansions and 1-particle reduced density matrix occupation numbers, as well as separation of electron correlation energy to dynamic and nondynamic components, which will enable fundamental understanding of FNA limits and decoupling of FN-bias from other bias sources of FNDMC. In addition, we plan large-scale screening of FNDMC accuracy in model systems covering various interaction regimes (weak, strong), method improvements, and applications to technologically important problems like stability and optical properties of 2D materials, semiconductors, and, catalysis on transition metals and their oxides. In addition to deep physical insights to the electron correlation in complex many-electron systems and limits of FNDMC methodology, the results of the project will enable rational usage and fine bias control of this method valuable for large systems.|
|Kind of project:||APVV - Všeobecná výzva ()|
|Department:||Advanced Technologies Research Institute (MTF)|
|Project status:||Not approved|
|Project start date :||01. 07. 2018|
|Project close date:||31. 12. 2021|
|Number of workers in the project:||5|
|Number of official workers in the project:||5|
The following table shows the researcher workers who have been assigned official roles in a project.
|Ing. Matúš Dubecký, PhD.||zodpovedný riešiteľ|
|Mária Rešetková, MBA||administratíva|
|RNDr. Andrej Antušek, PhD.||spoluriešiteľ|
|doc. Ing. Stanislav Minárik, PhD.||spoluriešiteľ|
|RNDr. Martin Šulka, PhD.||spoluriešiteľ|