Accelerated calibration of semi-analytic galaxy formation models

Thu, 01 Jan 2026 00:00:00 +0000

Understanding how galaxies form is one of the major challenges in astrophysics. Researchers often use “semi-analytic models” — simplified but physically motivated simulations — to explore how processes such as gas cooling, star formation, and feedback from stars and black holes shape galaxy populations.

These models contain many uncertain parameters that must be calibrated so that the simulated galaxies resemble real ones observed in the Universe. Unfortunately, traditional calibration methods require running very large simulations, which can be computationally expensive and slow.

In this work we introduce a faster way to calibrate these models. Instead of matching the full distribution of galaxy stellar masses directly, we use the relationship between the stellar mass of a galaxy and the mass of the dark matter halo that hosts it. This allows us to evaluate each model using only a small number of simulated galaxies rather than thousands.

Using this approach with the galaxy formation model, we show that it can reproduce key observational measurements while allowing rapid exploration of model assumptions. This makes it much easier to test different physical prescriptions and identify promising models for more detailed simulations.

Galacticus | Andrew Robertson

Accelerated calibration of semi-analytic galaxy formation models