A1009
Title: Deep impulse control
Authors: Bowen Jia - The Chinese University of Hong Kong (Hong Kong) [presenting]
Hoi Ying Wong - The Chinese University of Hong Kong (Hong Kong)
Abstract: A novel deep learning framework is developed to estimate the optimal control policy for the impulse control problem. Using deep neural networks, this numerical method allows for a general class of stochastic processes and consideration of uncontrollable co-integrated processes. The method also applies to high-dimensional cases for controllable and uncontrollable stochastic processes. This method can solve the optimal policy for a wide range of impulse control problems, such as irreversible reinsurance, interest rate intervention and stochastic inventory control.
