Estimating halo merging timescales through emulation-based models

Cosmic structure formation proceeds in a bottom-up manner, whereby small structures form first and then coalesce together to form more massive ones. This process is largely driven by dynamical friction, which is the result of a lagging wake of mass behind the least massive object, inducing a dragging force that removes its orbital energy and angular momentum. The efficiency of this process, and hence the timescale on which mergers occur, depends on the orbital and structural properties of the objects that are involved. Previous models used to predict merging timescales are based on analytical arguments or formulas calibrated to cosmological simulations, which have resulted in varying degrees of success. In this project, we will leverage targeted simulations that sample a broad range of possible parameter combinations – e.g. orbital energy, eccentricity, relative masses – together with machine learning techniques to explore how well previous models do, what role does numerical resolution play, and whether emulation-based models fare better than traditional approaches.