Delta-Gamma-Like Hedging with Transaction Cost under Reinforcement Learning Technique

Option hedging is critical in financial risk management. The traditional methods to determine the hedging position require assumptions of a frictionless market and continuous hedging. In this article, these two assumptions are removed, and a hedging strategy based on the reinforcement learning technique is proposed. This new strategy maximizes the expectation of the present value of accounting and realized profits of the hedging portfolio while limiting the sensitivity of the hedging position to changes in the underlying asset. The performance of this method is tested on option trading data (from 2004 to 2020) for the Standard and Poor’s (S&P) 500, S&P 100, and Dow Jones Industrial Average.