Generalized Policy Improvement Algorithms with Theoretically Supported Sample Reuse
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2022
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916434026692608 |
|---|---|
| author | Queeney, James Paschalidis, Ioannis Ch. Cassandras, Christos G. |
| author_facet | Queeney, James Paschalidis, Ioannis Ch. Cassandras, Christos G. |
| contents | We develop a new class of model-free deep reinforcement learning algorithms for data-driven, learning-based control. Our Generalized Policy Improvement algorithms combine the policy improvement guarantees of on-policy methods with the efficiency of sample reuse, addressing a trade-off between two important deployment requirements for real-world control: (i) practical performance guarantees and (ii) data efficiency. We demonstrate the benefits of this new class of algorithms through extensive experimental analysis on a broad range of simulated control tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2206_13714 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Generalized Policy Improvement Algorithms with Theoretically Supported Sample Reuse Queeney, James Paschalidis, Ioannis Ch. Cassandras, Christos G. Machine Learning Artificial Intelligence We develop a new class of model-free deep reinforcement learning algorithms for data-driven, learning-based control. Our Generalized Policy Improvement algorithms combine the policy improvement guarantees of on-policy methods with the efficiency of sample reuse, addressing a trade-off between two important deployment requirements for real-world control: (i) practical performance guarantees and (ii) data efficiency. We demonstrate the benefits of this new class of algorithms through extensive experimental analysis on a broad range of simulated control tasks. |
| title | Generalized Policy Improvement Algorithms with Theoretically Supported Sample Reuse |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2206.13714 |