<div class="csl-bib-body">
<div class="csl-entry">Mischek, F., & Musliu, N. (2022, July 0). <i>Investigating Hyper-heuristics for Real-World Test Laboratory Scheduling</i> [Conference Presentation]. EURO 2022, Espoo, Finland.</div>
</div>
-
dc.identifier.uri
http://hdl.handle.net/20.500.12708/136987
-
dc.description.abstract
Selection hyper-heuristics are a class of high-level problem solving methods that operate on a set of problem-specific low-level operators, instead of directly over a solution space. This allows them to be adaptive and problem-independent, even for previously unknown problem domains. One such domain is the Test Laboratory Scheduling Problem (TLSP), which is an extension of the well-known RCPSP. In the TLSP, solvers have to produce a schedule by first grouping tasks into jobs and then assigning a mode, time slot, and resources to the jobs. We have modeled the TLSP as a problem domain for selection hyper-heuristics, using the HyFlex framework, which has become the de-facto
standard in this area. This includes defining a solution representation and evaluation function, as well as a diverse portfolio of low-level operators of different types, including large neighborhood operators. Using a newly developed and problem-independent hyper-heuristic based on reinforcement learning, we were able to find high-quality solutions for the TLSP that are competitive with the current state of the art and could even improve the results for some instances. In addition, our hyper-heuristics produce good results on both small and large instances, compared to the existing problem-specific approaches, whose performance depends a lot on the instance size. We could further show the generality of our approach by achieving high scores on the original problem domains contained in HyFlex, which contain a selection of the most well-known NP-complete optimization problems.
en
dc.language.iso
en
-
dc.subject
TLSP
en
dc.subject
RCPSP
en
dc.subject
Hyper-heuristics
en
dc.title
Investigating Hyper-heuristics for Real-World Test Laboratory Scheduling
en
dc.type
Presentation
en
dc.type
Vortrag
de
dc.type.category
Conference Presentation
-
tuw.researchTopic.id
I1
-
tuw.researchTopic.id
C5
-
tuw.researchTopic.name
Logic and Computation
-
tuw.researchTopic.name
Computer Science Foundations
-
tuw.researchTopic.value
90
-
tuw.researchTopic.value
10
-
tuw.linking
https://euro2022espoo.com/conference-programme/
-
tuw.publication.orgunit
E192-02 - Forschungsbereich Databases and Artificial Intelligence
-
tuw.author.orcid
0000-0003-1166-3881
-
tuw.author.orcid
0000-0002-3992-8637
-
tuw.event.name
EURO 2022
en
tuw.event.startdate
03-07-2022
-
tuw.event.enddate
06-07-2022
-
tuw.event.online
On Site
-
tuw.event.type
Event for scientific audience
-
tuw.event.place
Espoo
-
tuw.event.country
FI
-
tuw.event.presenter
Mischek, Florian
-
tuw.event.track
Multi Track
-
wb.sciencebranch
Informatik
-
wb.sciencebranch.oefos
1020
-
wb.sciencebranch.value
100
-
item.languageiso639-1
en
-
item.openairetype
conference paper not in proceedings
-
item.grantfulltext
none
-
item.fulltext
no Fulltext
-
item.cerifentitytype
Publications
-
item.openairecristype
http://purl.org/coar/resource_type/c_18cp
-
crisitem.author.dept
E192-02 - Forschungsbereich Databases and Artificial Intelligence
-
crisitem.author.dept
E192-02 - Forschungsbereich Databases and Artificial Intelligence
