---
res:
bibo_abstract:
- "Algorithm selection (AS) deals with the automatic selection of an algorithm\r\nfrom
a fixed set of candidate algorithms most suitable for a specific instance\r\nof
an algorithmic problem class, where \"suitability\" often refers to an\r\nalgorithm's
runtime. Due to possibly extremely long runtimes of candidate\r\nalgorithms, training
data for algorithm selection models is usually generated\r\nunder time constraints
in the sense that not all algorithms are run to\r\ncompletion on all instances.
Thus, training data usually comprises censored\r\ninformation, as the true runtime
of algorithms timed out remains unknown.\r\nHowever, many standard AS approaches
are not able to handle such information in\r\na proper way. On the other side,
survival analysis (SA) naturally supports\r\ncensored data and offers appropriate
ways to use such data for learning\r\ndistributional models of algorithm runtime,
as we demonstrate in this work. We\r\nleverage such models as a basis of a sophisticated
decision-theoretic approach\r\nto algorithm selection, which we dub Run2Survive.
Moreover, taking advantage of\r\na framework of this kind, we advocate a risk-averse
approach to algorithm\r\nselection, in which the avoidance of a timeout is given
high priority. In an\r\nextensive experimental study with the standard benchmark
ASlib, our approach is\r\nshown to be highly competitive and in many cases even
superior to\r\nstate-of-the-art AS approaches.@eng"
bibo_authorlist:
- foaf_Person:
foaf_givenName: Alexander
foaf_name: Tornede, Alexander
foaf_surname: Tornede
foaf_workInfoHomepage: http://www.librecat.org/personId=38209
- foaf_Person:
foaf_givenName: Marcel Dominik
foaf_name: Wever, Marcel Dominik
foaf_surname: Wever
foaf_workInfoHomepage: http://www.librecat.org/personId=33176
orcid: ' https://orcid.org/0000-0001-9782-6818'
- foaf_Person:
foaf_givenName: Stefan
foaf_name: Werner, Stefan
foaf_surname: Werner
- foaf_Person:
foaf_givenName: Felix
foaf_name: Mohr, Felix
foaf_surname: Mohr
- foaf_Person:
foaf_givenName: Eyke
foaf_name: HÃ¼llermeier, Eyke
foaf_surname: HÃ¼llermeier
foaf_workInfoHomepage: http://www.librecat.org/personId=48129
dct_date: 2020^xs_gYear
dct_language: eng
dct_title: 'Run2Survive: A Decision-theoretic Approach to Algorithm Selection based
on Survival Analysis@'
...