Job Offers - Research Positions
Reproducibility of thermodynamic simulations
Reproducibility is of fundamental importance for science. In laboratory experiments, reproducibility is a commonly accepted goal and efforts are made to fulfill it. In contrast, in computer experiments, it is often assumed that results would be exactly reproducible; a standpoint which, however, has been shown to be naive. The complexity of computer simulations influences the reproducibility in many ways, e.g. by the physical and numerical algorithms, hardware and software, as well as the social organization of the simulation. These issues will be addressed and analyzed in this project using thermodynamic simulations as an example. The focus is on molecular simulations and equations of state. The goal is to develop methods to improve the reproducibility of thermodynamic simulations, based on a survey of the current practice. Simulation tasks, which are used as examples here, will be taken from ongoing projects, so that the results from the project will be used directly in practice.