In this work, we propose a novel approach for reinforcement learning driven by evolutionary computation. Our algorithm, dubbed as Evolutionary-Driven Reinforcement Learning (evo-RL), embeds the reinforcement learning algorithm in an evolutionary cycle, where we distinctly differentiate between purely evolvable (instinctive) behaviour versus purely learnable behaviour. Furthermore, we propose that this distinction is decided by the evolutionary process, thus allowing evo-RL to be adaptive to different environments. In addition, evo-RL facilitates learning on environments with rewardless states, which makes it more suited for real-world problems with incomplete information. To show that evo-RL leads to state-of-the-art performance, we present the performance of different state-of-the-art reinforcement learning algorithms when operating within evo-RL and compare it with the case when these same algorithms are executed independently. Results show that reinforcement learning algorithms embedded within our evo-RL approach significantly outperform the stand-alone versions of the same RL algorithms on OpenAI Gym control problems with rewardless states constrained by the same computational budget.
EVO-RL: Evolutionary-Driven Reinforcement Learning / Hallawa, Ahmed; Born, Thorsten; Schmeink, Anke; Dartmann, Guido; Peine, Arne; Martin, Lukas; Iacca, Giovanni; Eiben, A. E.; Ascheid, Gerd. - (2021), pp. 153-154. (Intervento presentato al convegno 2021 Genetic and Evolutionary Computation Conference, GECCO 2021 tenutosi a Lille, France, nel 10th-14th July 2021) [10.1145/3449726.3459475].
EVO-RL: Evolutionary-Driven Reinforcement Learning
Iacca, Giovanni;
2021-01-01
Abstract
In this work, we propose a novel approach for reinforcement learning driven by evolutionary computation. Our algorithm, dubbed as Evolutionary-Driven Reinforcement Learning (evo-RL), embeds the reinforcement learning algorithm in an evolutionary cycle, where we distinctly differentiate between purely evolvable (instinctive) behaviour versus purely learnable behaviour. Furthermore, we propose that this distinction is decided by the evolutionary process, thus allowing evo-RL to be adaptive to different environments. In addition, evo-RL facilitates learning on environments with rewardless states, which makes it more suited for real-world problems with incomplete information. To show that evo-RL leads to state-of-the-art performance, we present the performance of different state-of-the-art reinforcement learning algorithms when operating within evo-RL and compare it with the case when these same algorithms are executed independently. Results show that reinforcement learning algorithms embedded within our evo-RL approach significantly outperform the stand-alone versions of the same RL algorithms on OpenAI Gym control problems with rewardless states constrained by the same computational budget.File | Dimensione | Formato | |
---|---|---|---|
2007.04725.pdf
accesso aperto
Tipologia:
Pre-print non referato (Non-refereed preprint)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
5.74 MB
Formato
Adobe PDF
|
5.74 MB | Adobe PDF | Visualizza/Apri |
3449726.3459475.pdf
Solo gestori archivio
Tipologia:
Versione editoriale (Publisher’s layout)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
382.3 kB
Formato
Adobe PDF
|
382.3 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione