A1671
Title: Dynamic decision making with reinforcement learning
Authors: Jozef Barunik - UTIA AV CR vvi (Czech Republic)
Lukas Vacha - Institute of Information Theory and Automation of the CAS (Czech Republic) [presenting]
Abstract: A solution is proposed to a general class of models under uncertainty with agents having quantile preferences and limited information processing capacity. We demonstrate our reinforcement learning approach on a simple example where the agent acquires an optimal amount of information. The agent has a limited amount of attention since the information he obtains is costly. Our method can be further extended to more complicated high-dimensional problems, where an analytical solution is impossible to obtain, whereas our reinforcement learning approach makes this task computationally feasible.
