At the beginning of the 21st century, a new phase state of strongly interacting matter was
established, known as the quark-gluon plasma (QGP) [1]. To study the formation of the QGP in
collisions of heavy nuclei, specific equation of state (EoS) is typically employed. Lattice Quantum
Chromodynamics (LQCD) is an effective tool for studying the thermochemical evolution of strongly
interacting matter, but it encounters difficulties for non-zero baryonic potentials. In light of this,
various holographic models based on the well-known AdS/CFT duality have been proposed to obtain
EoS for the QGP.
In the presented report, which continues our previous work [2], [3], [4], a calibration method is
proposed for the holographic EoS from I.Ya. Aref’eva’s theoretical group to study the properties of
QGP within the framework of relativistic hydrodynamics. To solve a problem with a wide range of
unknown parameters using symbolic regression, a custom reinforcement learning algorithm is applied.
Furthermore, we discuss specific modifications to the approach to achieve better agreement with the
hadron gas EoS predictions for a small temperature range and taking into account several hadronic
quantum numbers