Reinforcement Learning for Model Problems of Optimal Control

S. S. Semenov; Семенов С. С.; V. I. Tsurkov; Цурков В. И.

doi:10.31857/S0002338823030125

Reinforcement Learning for Model Problems of Optimal Control

Authors: Semenov S.S.¹, Tsurkov V.I.²
Affiliations:
1. Moscow Institute of Physics and Technology, 141701, Dolgoprudny, Moscow Oblast, Russia
2. Federal Research Center “Computer Science and Control,” Russian Academy of Sciences, 119333, Moscow, Russia
Issue: No 3 (2023)
Pages: 76-89
Section: ARTIFICIAL INTELLIGENCE
URL: https://ruspoj.com/0002-3388/article/view/676487
DOI: https://doi.org/10.31857/S0002338823030125
EDN: https://elibrary.ru/EVAFAM
ID: 676487

Cite item

Full Text

Open Access
Restricted Access

Access granted
Restricted Access

Subscription or Fee Access

Abstract

The functionals of dynamic systems of various types are optimized using modern methods of reinforcement learning. The linear resource allocation problem, as well as the optimal consumption problem and its stochastic modifications are considered. In the reinforcement learning strategy gradient methods are used.

About the authors

S. S. Semenov

Moscow Institute of Physics and Technology, 141701, Dolgoprudny, Moscow Oblast, Russia

Email: semenov.ss@phystech.edu
Россия, МО, Долгопрудный

V. I. Tsurkov

Federal Research Center “Computer Science and Control,” Russian Academy of Sciences, 119333, Moscow, Russia

Author for correspondence.
Email: tsur@ccas.ru
Россия, Москва

References

Sewak M. Deterministic Policy Gradient and the DDPG: Deterministic-Policy-Gradient-Based Approaches. 2019.
Schulman J. Trust Region Policy Optimization. 2015. https://arxiv.org/abs/1502.05477.
Haarnoja T. Soft Actor-Critic: Off-Policy Maximum Entropy Deep Rein-forcement Learning with a Stochastic Actor. 2018. https://arxiv.org/abs/1801.01290.
Huang S. A2C is a special case of PPO. 2022. https://arxiv.org/abs/2205.09123.
Schulman J. Proximal Policy Optimization Algorithms. 2017. https://arxiv.org/abs/1707.06347.
Zhang L. Penalized Proximal Policy Optimization for Safe Reinforcement Learning. 2022. https://arxiv.org/abs/2205.11814.
Chen X. The Sufficiency of Off-policyness: PPO is insufficient according to an Off-policy Measure. 2022. https://arxiv.org/abs/2205.10047.
Ghosh A. Provably Efficient Model-Free Constrained RL with Linear Function Approximation. 2022. https://arxiv.org/abs/2206.11889.
Song Z. Safe-FinRL: A Low Bias and Variance Deep Reinforcement Learning Implementation for High-Freq Stock Trading. 2022. https://arxiv.org/abs/2206.05910.
Kaledin M. Variance Reduction for Policy-Gradient Methods via Empirical Variance Minimization. 2022. https://arxiv.org/abs/2206.06827.
Luo Q. Finite-Time Analysis of Fully Decentralized Single-Timescale Actor- Critic. 2022. https://arxiv.org/abs/2206.05733.
Deka A. ARC – Actor Residual Critic for Adversarial Imitation Learning. 2022. https://arxiv.org/abs/2206.02095.
Цурков В.И. Динамические задачи большой размерности. М.: Наука, 1988. 287 с.
Бекларян Л.А., Флёрова А.Ю., Жукова А.А. Методы оптимального управления: учеб. пособие. М.: Наука, 2018.
Оксендаль Б. Стохастические дифференциальные уравнения. Введение в теорию и приложеия. М.: Мир, 2003.
Понтрягин Л.С. Принцип максимума в оптимальном управлении. М.: Наука, 2004.