Speaker
Description
Rare-event search experiments are continuously extending their sensitivities to unprecedented levels. Achieving these feats requires increasingly small backgrounds, partly as a result of improved shielding schemes and deep underground laboratories that can suppress external backgrounds by several orders of magnitude. Detailed detector and shielding simulations are required to attain a good understanding of experimental backgrounds, and the achieved sensitivity. With backgrounds now frequently below 0.01 counts per kg of target per keVee of energy, said simulations require increasingly more computing resources. Event biasing is often applied to mitigate this issue, particularly in shielding simulations, yet there is a lack of studies providing systematic guidance on how to best optimise biasing techniques for statistical precision and CPU-time, limiting the obtained benefit. Such an optimisation study for the importance-splitting biasing technique implemented in GEANT4 will be discussed, focused on balancing statistical precision with simulation CPU-time. This may also result in reduced computing-related CO2$_{\rm e}$ emissions, which is also discussed.