Skip to main content

Theory and Algorithms for Adaptive Particle Simulation

Final Report Summary - ADAPT (Theory and Algorithms for Adaptive Particle Simulation)

The general goal of the project was to develop a theory, as well as associated algorithms, for adaptively restrained (AR) particle simulation. AR simulation consists in adaptively freezing and unfreezing particles to speed up simulation, and has applications in nanoscience (biology, physics, chemistry, etc.), but also in other fields benefiting from particle simulations (e.g. computational fluid simulations, computer graphics, etc.).

The project was organized into six main tasks related to the AR method:

Task 1 – Defining adaptive Hamiltonians

Task 2 – Developing algorithms for adaptive potential update – classical cases

Task 3 – Developing algorithms for adaptive potential update – quantum cases

Task 4 – Developing adaptive particle simulation algorithms on parallel hardware

Task 5 – Developing algorithms for accelerated sampling

Task 6 – Software development of SAMSON

The ERC grant has enabled us to establish a theoretical and computational framework around the AR simulation methodology, and to demonstrate its practicality and benefit.

Furthermore, we have greatly extended SAMSON, the software platform for computational nanoscience developed in the group, in order to make it possible to integrate all algorithms developed during the project. SAMSON was released in March 2015 at and appears to be attracting a lot of attention (more that 1200 beta testers in October 2017).