The results demonstrate that under transaction costs, the DRL agent significantly outperforms the BS Delta strategy, showcasing a higher mean final P&L and a lower standard deviation in performance.
When assessing a scenario without transaction costs, both the DRL agent and the BS Delta strategy yield near-zero means, indicating effective hedging performance, yet the DRL strategy consistently provides a higher mean final P&L.
With a 3% transaction cost, the DRL agent exhibits a higher mean final P&L across multiple options, outperforming the BS Delta strategy in numerous symbol and maturity combinations.
The findings of the study highlight the advantages of DRL strategies, particularly in environments with transaction costs, emphasizing its capability to enhance hedging strategies through superior performance metrics.
#deep-reinforcement-learning #hedging-strategies #financial-markets #stochastic-volatility #transaction-costs
Collection
[
|
...
]