EUVIMEDEUVIMED entdecken
Crossref

Developing Monte Carlo Simulator of Reinforcement Learning Type

Georgi Tsochev

Problems of Engineering Cybernetics and Robotics · 2020

DOI: 10.7546/pecr.73.20.04 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Problems of Engineering Cybernetics and Robotics, 2020. Verfasst von Georgi Tsochev.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Monte Carlo methods are a way to solve the reinforcement learning problem based on average test results. To ensure that well-defined results are available, Monte Carlo methods are used only for episodic tasks. The Monte Carlo term is often used more widely in any valuation method whose operation involves significant participation on a random basis. Here it is specifically used for methods based on the average of full results (as opposed to methods that are learned from incomplete results). The paper describes a simulator for estimating raindrops in a specific area using the package matlib. Keywords: Monte Carlo, reinforcement learning, simulation, matlib, python

Zitieren

APA

Tsochev, G. (2020). Developing Monte Carlo Simulator of Reinforcement Learning Type. Problems of Engineering Cybernetics and Robotics. https://doi.org/10.7546/pecr.73.20.04

Vancouver

Tsochev G. Developing Monte Carlo Simulator of Reinforcement Learning Type. Problems of Engineering Cybernetics and Robotics. 2020. doi:10.7546/pecr.73.20.04

BibTeX
@article{Tsochev2020,
  author = {Tsochev, Georgi},
  title = {Developing Monte Carlo Simulator of Reinforcement Learning Type},
  journal = {Problems of Engineering Cybernetics and Robotics},
  year = {2020},
  doi = {10.7546/pecr.73.20.04}
}
Kostenlos starten →