State constrained stochastic optimal control for continuous and hybrid dynamical systems using DFBSDE

被引：3

作者：

Dai, Bolun ^{[1
]}

Krishnamurthy, Prashanth ^{[1
]}

Papanicolaou, Andrew ^{[2
]}

Khorrami, Farshad ^{[1
]}

机构：

[1] NYU, Brooklyn, NY 11201 USA

[2] North Carolina State Univ, Raleigh, NC 27695 USA

来源：

AUTOMATICA | 2023年 / 155卷

基金：

美国国家科学基金会;

关键词：

Stochastic control; Optimal control; Forward and backward stochastic; differential equations; MPC;

D O I：

10.1016/j.automatica.2023.111146

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

We develop a computationally efficient learning-based forward-backward stochastic differential equa-tions (FBSDE) controller for both continuous and hybrid dynamical (HD) systems subject to stochas-tic noise and state constraints. Solutions to stochastic optimal control (SOC) problems satisfy the Hamilton-Jacobi-Bellman (HJB) equation. Using current FBSDE-based solutions, the optimal control can be obtained from the HJB equations using deep neural networks (e.g., long short-term memory (LSTM) networks). To ensure the learned controller respects the constraint boundaries, we enforce the state constraints using a soft penalty function. In addition to previous works, we adapt the deep FBSDE (DFBSDE) control framework to handle HD systems consisting of continuous dynamics and a deterministic discrete state change. We demonstrate our proposed algorithm in simulation on a continuous nonlinear system (cart-pole) and a hybrid nonlinear system (five-link biped).& COPY; 2023 Elsevier Ltd. All rights reserved.

引用

页数：7