Providing the optimal stock portfolio pattern through limiting random dominance and reducing absolute risk aversion

Stochastic programming has proven to be a powerful modeling tool when an accurate probabilistic description, ie the exact values ​​of system parameters and the specific probability distributions for random variables, is available. However, such information is rarely available in practice. In such a situation, there are two main ways to deal with ambiguities. One through-sample average (SAA) is also known as the Monte Carlo method, where the SAA is constructed from the expected value of the underlying performance using experimental data. The aim of this study was to provide an optimal stock portfolio model by limiting random dominance and reducing absolute risk aversion. The statistical sample of the stock exchange and the type of data collected from that time series are the changes and cumulative changes of the total index of the Tehran Stock Exchange in 1390 to 1400. The stock portfolio is based on information from 50 companies. First, these data are examined, then the desired algorithms are designed and the model is analyzed based on the necessary conditions and tests.Performance is evaluated based on Sharp, Trainer, Sortino, and Omega tests. In the following, the stock portfolio is determined based on the initial assumptions and is modeled and measured based on the companies' returns.