May 30, 2019 · This article discusses a new application of reinforcement learning: to the problem of hedging a portfolio of “over-the-counter” derivatives under under market frictions such as trading costs and liquidity constraints.
Deep hedging. We present a framework for hedging a portfolio of derivatives in the presence of market frictions such as transaction costs, liquidity constraints or risk limits using modern deep reinforcement machine learning methods.
Reinforcement Learning works by: Providing an opportunity or degree of freedom to enact a behavior - such as making decisions or choices. Providing contextual information about the environment and choices.
2 complimentary, easy-to-read blog posts to get started on deep reinforcement learning: the first one focuses on policy gradients, the second one focuses on deep Q-learning. Deep Reinforcement Learning: Pong from Pixels ( mirror ) by Andrej Karpathy (May 31, 2016). Deep Hedging. Article ... in the presence of market frictions such as transaction costs, liquidity constraints or risk limits using modern deep reinforcement machine learning methods. We discuss ... Dynamic Replication and Hedging: A Reinforcement Learning Approach Petter Kolm, New York University In this talk, we address the problem of how to optimally hedge an options book in a practical setting, where trading decisions are discrete and trading costs can be nonlinear and difficult to model. Deep Reinforcement Learning (DRL) Dynamic Programming not suitable in practice due to: Curse of Dimensionality Curse of Modeling So we solve the MDP with Deep Reinforcement Learning (DRL) The idea is to use real market data and real market frictions Developing realistic simulations to derive the optimal policy