Application of machine learning to study the effects of quadrupole deformation on the nucleus in heavy-ion collisions at intermediate energies
Abstract
The influence of quadrupole deformation on the projectile and target nucleus in $^{238}$U + $^{238}$U collisions at beam energies of 0.4, 1.0, and 1.5 GeV/nucleon is investigated using the ultra relativistic quantum molecular dynamics (UrQMD) model. The effect of quadrupole deformation is investigated using the LightGBM decision tree algorithm. LightGBM's ability to identify nuclear deformation from particle spectra is demonstrated on an event-by-event basis by learning the two-dimensional (transverse momentum and rapidity) distributions of free protons, charged fragments (mass number $A>$1), and charged $\pi$ mesons produced in $^{238}$U + $^{238}$U collisions with and without deformed initializations. Using the important characteristic analysis of LightGBM, the yield of charged fragments around target/projectile rapidities is found to be sensitive to the quadrupole deformation of both the projectile and the target.It can be used as a promising observable to probe quadrupole deformation with heavy-ion collision.
