In recent years, the expansion of financial markets and the increasing complexity of price behavior have made the use of intelligent methods in designing trading strategies a necessity. The gold futures market, as one of the highly volatile markets sensitive to macroeconomic variables, provides a suitable environment for developing and evaluating advanced algorithmic trading systems. In this study, a novel framework based on Deep Reinforcement Learning is proposed for designing an optimal algorithmic trading system in the gold futures market.
The proposed model combines deep neural networks with reinforcement learning algorithms such as Deep Q-Network (DQN) and Proximal Policy Optimization (PPO) to learn an optimal trading policy through continuous interaction with the market environment. In this framework, the market state includes variables such as price, trading volume, and technical indicators, while the reward function is designed to consider not only profitability but also risk in the decision-making process.
Preliminary simulation results show that the proposed model outperforms traditional technical analysis strategies such as Moving Average and MACD in terms of risk-adjusted return (Sharpe Ratio) and drawdown control. The findings indicate that deep reinforcement learning can serve as an effective tool for designing intelligent trading systems in complex and nonlinear financial markets.