Search academic texts

Information × Registration Number 0225U001586, (0123U102998) , R & D reports Title Modeling systems of strongly interacting and dark matter popup.stage_title Статистичний аналіз космологічних моделювань Всесвіту та симуляцій зіткнень ядер золота за релятивістських енергій з використанням нейронних мереж. Head Rudakovskyi Anton V., Кандидат фізико-математичних наук Registration Date 03-02-2025 Organization М.М.Bogolyubov Institute of Theoretical Physics of the National Academy of Sciences of Ukraine popup.description1 Assessing the possibility of constraining the amplitude of density perturbations of dark matter in the early Universe using future radio astronomical observations. Modeling such observations of neutral hydrogen radiation at the 21-cm wavelength. Applying Bayesian analysis to the simulated observations. Modeling nucleus-nucleus collisions at the Relativistic Heavy Ion Collider. Exploring new properties of strongly interacting matter at high baryon charge densities. Estimating the shear viscosity and thermalization rate of quark-gluon plasma. Comparing experimental results with model simulations and searching for anomalies in the phase diagram of quantum chromodynamics that could indicate a phase transition and critical point. popup.description2 The object of the study is the Early Universe, dark matter, quark-gluon plasma, and heavy-ion collisions. The aim of the project is to investigate the influence of dark matter on the formation of cosmological structures, as well as to model processes in collider experiments. The feasibility of determining the parameters Omega_m and sigma_8 using neural networks based on CAMELS simulations has been explored. The analysis utilized the monopole and quadrupole components of galaxy-galaxy and galaxy-matter correlation functions. The network demonstrates the ability to estimate these parameters with an accuracy of 5-6%, surpassing the results of similar studies. A method for combining data from different simulations while accounting for cosmic variance effects is proposed. This approach enhances the resemblance of simulated data to real Universe observations. The integrated hydrodynamic model has been modified. A C++ code has been developed to compute the energy-momentum tensor of the non-equilibrium hadron gas in Milne coordinates. The code has been integrated into the main model using Bash and Python scripts. Simulations of nuclear-nuclear collisions for RHIC BES experiments have been performed based on this model. The analysis demonstrates that neural networks can effectively recover optimal model parameters by analyzing the spectra of charged hadrons. The results obtained confirm the effectiveness of using machine learning and numerical modeling to address fundamental problems in cosmology and high-energy physics. Product Description popup.authors Adzhymambetov Musfer D. popup.nrat_date 2025-02-03 Close