<div class="csl-bib-body">
<div class="csl-entry">Elgeti, S., Wolff, D., Fricke, C. D., & Kemmerling, M. (2023, April 26). <i>Reinforcement Learning For Spline-Based Shape Optimization Of Flow Channels In Profile Extrusion Dies</i> [Conference Presentation]. 22nd Computational Fluids Conference CFC 2023, Cannes, France. http://hdl.handle.net/20.500.12708/191065</div>
</div>
-
dc.identifier.uri
http://hdl.handle.net/20.500.12708/191065
-
dc.description.abstract
In this contribution, we investigate the feasibility of Reinforcement Learning (RL) as learning-based algorithm for shape optimization. Reinforcement Learning is based on a trial-and-error interaction of an agent with an environment. For each action, the agent is informed about a reward and the subsequent state of the environment, but there is no information about long-term interests as classical optimization algorithms would provide. While not necessarily superior to classical, e.g., gradient-based, optimization algorithms for one single optimization problem, Reinforcement Learning techniques are expected to perform especially well on similar optimization tasks, since the agent learns a general strategy for solving a problem instead of just concentrating on the solution of a single problem.
We introduce an application case for RL-based shape optimization, in which the agent can modify the geometry of a flow channel and the resulting computational mesh through a spline-based deformation method known as Free Form Deformation by manipulating the control point coordinates of the transfor- mation spline. Every action of the agent requires computing the new state of the environment by performing a high-fidelity Finite Element Method (FEM) simulation in the modified geometry.
As a first step, a modular python framework for driving the interaction be- tween the RL library stable- baselines3 [1] and the in-house FEM solver has been created and tested on a 2D example. Based on this, we will present different ways to improve the performance of the RL framework, e.g, by shaping the reward function, employing a multi-environment approach as well as comparing on- and off-policy algorithms, where the latter can be expected to be more sample-efficient, all of them paving the way towards optimizing more complex, industrial geometries. Another contribution will be the training of a custom flow feature extractor, which we expect to improve the generalization properties of the agent so that it can be reused for transfer learning tasks which becomes especially useful for addressing multi- query scenarios.
en
dc.language.iso
en
-
dc.subject
reinforcement learning
en
dc.subject
shape optimization
en
dc.title
Reinforcement Learning For Spline-Based Shape Optimization Of Flow Channels In Profile Extrusion Dies
en
dc.type
Presentation
en
dc.type
Vortrag
de
dc.contributor.affiliation
RWTH Aachen University, Germany
-
dc.contributor.affiliation
RWTH Aachen University, Germany
-
dc.type.category
Conference Presentation
-
tuw.researchTopic.id
C4
-
tuw.researchTopic.id
C5
-
tuw.researchTopic.id
C6
-
tuw.researchTopic.name
Mathematical and Algorithmic Foundations
-
tuw.researchTopic.name
Computer Science Foundations
-
tuw.researchTopic.name
Modeling and Simulation
-
tuw.researchTopic.value
50
-
tuw.researchTopic.value
30
-
tuw.researchTopic.value
20
-
tuw.publication.orgunit
E317-01-1 - Forschungsgruppe Numerische Analyse- und Designmethoden
-
tuw.publication.orgunit
E317-01-2 - Forschungsgruppe Struktur- und Werkstoffsimulation
-
tuw.author.orcid
0000-0002-4474-1666
-
tuw.author.orcid
0000-0002-9495-2285
-
tuw.author.orcid
0000-0003-0141-2050
-
tuw.event.name
22nd Computational Fluids Conference CFC 2023
en
tuw.event.startdate
25-04-2023
-
tuw.event.enddate
28-04-2023
-
tuw.event.online
On Site
-
tuw.event.type
Event for scientific audience
-
tuw.event.place
Cannes
-
tuw.event.country
FR
-
tuw.event.presenter
Elgeti, Stefanie
-
tuw.event.track
Multi Track
-
wb.sciencebranch
Maschinenbau
-
wb.sciencebranch
Informatik
-
wb.sciencebranch
Mathematik
-
wb.sciencebranch.oefos
2030
-
wb.sciencebranch.oefos
1020
-
wb.sciencebranch.oefos
1010
-
wb.sciencebranch.value
40
-
wb.sciencebranch.value
40
-
wb.sciencebranch.value
20
-
item.fulltext
no Fulltext
-
item.openairecristype
http://purl.org/coar/resource_type/c_18cp
-
item.languageiso639-1
en
-
item.cerifentitytype
Publications
-
item.openairetype
conference paper not in proceedings
-
item.grantfulltext
none
-
crisitem.author.dept
E317-01 - Forschungsbereich Leichtbau
-
crisitem.author.dept
E317-01 - Forschungsbereich Leichtbau
-
crisitem.author.dept
E317-01-2 - Forschungsgruppe Struktur- und Werkstoffsimulation
-
crisitem.author.dept
RWTH Aachen University
-
crisitem.author.orcid
0000-0002-4474-1666
-
crisitem.author.orcid
0000-0002-9495-2285
-
crisitem.author.orcid
0000-0003-0141-2050
-
crisitem.author.parentorg
E317 - Institut für Leichtbau und Struktur-Biomechanik
-
crisitem.author.parentorg
E317 - Institut für Leichtbau und Struktur-Biomechanik