Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Risk Aware Reinforcement Learning with Safety Layer

 
: Tuna, Ongun
: Buss, Martin; Leibold, Marion; Schmoeller da Roza, Felippe

München, 2021, 69 pp.
München, TU, Master Thesis, 2021
Bayerisches Staatsministerium für Wirtschaft, Landesentwicklung und Energie StMWi

IKS-Ausbauprojekt
English
Master Thesis
Fraunhofer IKS ()
reinforcement learning; RL; safety; risk; safety layer; distributional deep deterministic policy gradients; variance constrained policy gradients

Abstract
This thesis aims to analyze the effects of combining two different safe reinforcement learning algorithms to cover the shortcomings of each algorithms. Firstly, a safety layer algorithm which corrects the actions leading to error states is implemented. Safety layer is combined with two different risk sensitive reinforcement learning algorithms: a variance constrained deep deterministic policy gradient algorithm and a risk sensitive distributional deep deterministic policy gradients algorithm. The results are evaluated by comparing rewards, episode lengths, action corrections and variance of the returns provided by vanilla deep deterministic policy gradients algorithms and risk-aware deep deterministic policy gradients algorithms combined with safety layers.

: http://publica.fraunhofer.de/documents/N-642557.html