Determining the Optimal Agent Based on Profitability in the Financial Market Using Reinforcement Learning

In recent years, financial markets have drawn significant attention from investors and researchers due to their substantial profitability and relatively predictable patterns. The primary objective of trading agents in these markets is to maximize income while minimizing potential losses. For investment companies, identifying the optimal trading agent is crucial to ensuring profitability and guiding future financial strategies. This study addresses this challenge by designing and evaluating five automated strategic agents, each of which operates on various trading symbols. The performance of these agents is assessed based on their trading statements, with profitability serving as an important criterion. Using the epsilon-Greedy policy, this research determines the agent that consistently executes optimal actions over the long term. The findings of this study provide valuable insights into leveraging Reinforcement Learning (RL) techniques for optimizing decision-making processes in financial markets, enabling the design of more effective and adaptive investment strategies in dynamic and complex environments.