A modular framework for stabilizing deep reinforcement learning control
In Proceedings of the 22nd IFAC World Congress (To Appear),
Nathan P. Lawrence, Philip D. Loewen, Shuyuan Wang, Michael G. Forbes, R. Bhushan Gopaluni
[PDF] [arXiv]
Click to enlarge image.
Abstract
We propose a framework for the design of feedback controllers that combines the optimization-driven and model-free advantages of deep reinforcement learning with the stability guarantees provided by using the Youla-Kuˇcera parameterization to define the search domain. Recent advances in behavioral systems allow us to construct a data-driven internal model; this enables an alternative realization of the Youla-Kuˇcera parameterization based entirely on input-output exploration data. Using a neural network to express a parameterized set of nonlinear stable operators enables seamless integration with standard deep learning libraries. We demonstrate the approach on a realistic simulation of a two-tank system.
Read or Download: PDF