Recent research on automated trading algorithms has focused on evaluating their effectiveness on various financial markets and comparing their performance with that of other trading methods. This paper proposes an innovative framework for the simulation of trading algorithms with the purpose of supporting the development of automated decision systems for traders using agent-based modeling and reinforcement learning methods in a market context. The obtained results demonstrate that a trading agent can learn and optimize the trading strategies, even if performance variations were observed in relation to environmental changes. The success of the trading agent depends on the commitment of trading partners and the implementation of risk management in correlation with the market norms. The proposed model provides a valuable platform for further studies on learning behavior in trading based on the decisions of a real trader.
Artificial neural networks, Multi‐agent systems, Transaction databases, Reinforcement learning methods.
Doina BANCIU, Anca-Gabriela PETRESCU, Ionica ONCIOIU, Marius PETRESCU, Florentina-Raluca BÎLCAN, Adrian-Ionuț GHIBANU, "The Simulation Framework for Automated Trading Algorithms on Capital Markets", Studies in Informatics and Control, ISSN 1220-1766, vol. 33(4), pp. 51-58, 2024. https://doi.org/10.24846/v33i4y202405
All articles published by Studies in Informatics and Control (SIC) are published in accordance with the terms and conditions of the Creative Commons Attribution License (CC BY-NC-SA 4.0).