{"citation":{"chicago":"Seiler, Moritz, Janina Pohl, Jakob Bossek, Pascal Kerschke, and Heike Trautmann. “Deep Learning as a Competitive Feature-Free Approach for Automated Algorithm Selection on the Traveling Salesperson Problem.” In <i>Parallel Problem Solving from {Nature} (PPSN XVI)</i>, 48–64. Berlin, Heidelberg: Springer-Verlag, 2020. <a href=\"https://doi.org/10.1007/978-3-030-58112-1_4\">https://doi.org/10.1007/978-3-030-58112-1_4</a>.","short":"M. Seiler, J. Pohl, J. Bossek, P. Kerschke, H. Trautmann, in: Parallel Problem Solving from {Nature} (PPSN XVI), Springer-Verlag, Berlin, Heidelberg, 2020, pp. 48–64.","mla":"Seiler, Moritz, et al. “Deep Learning as a Competitive Feature-Free Approach for Automated Algorithm Selection on the Traveling Salesperson Problem.” <i>Parallel Problem Solving from {Nature} (PPSN XVI)</i>, Springer-Verlag, 2020, pp. 48–64, doi:<a href=\"https://doi.org/10.1007/978-3-030-58112-1_4\">10.1007/978-3-030-58112-1_4</a>.","ama":"Seiler M, Pohl J, Bossek J, Kerschke P, Trautmann H. Deep Learning as a Competitive Feature-Free Approach for Automated Algorithm Selection on the Traveling Salesperson Problem. In: <i>Parallel Problem Solving from {Nature} (PPSN XVI)</i>. Springer-Verlag; 2020:48–64. doi:<a href=\"https://doi.org/10.1007/978-3-030-58112-1_4\">10.1007/978-3-030-58112-1_4</a>","ieee":"M. Seiler, J. Pohl, J. Bossek, P. Kerschke, and H. Trautmann, “Deep Learning as a Competitive Feature-Free Approach for Automated Algorithm Selection on the Traveling Salesperson Problem,” in <i>Parallel Problem Solving from {Nature} (PPSN XVI)</i>, 2020, pp. 48–64, doi: <a href=\"https://doi.org/10.1007/978-3-030-58112-1_4\">10.1007/978-3-030-58112-1_4</a>.","apa":"Seiler, M., Pohl, J., Bossek, J., Kerschke, P., &#38; Trautmann, H. (2020). Deep Learning as a Competitive Feature-Free Approach for Automated Algorithm Selection on the Traveling Salesperson Problem. <i>Parallel Problem Solving from {Nature} (PPSN XVI)</i>, 48–64. <a href=\"https://doi.org/10.1007/978-3-030-58112-1_4\">https://doi.org/10.1007/978-3-030-58112-1_4</a>","bibtex":"@inproceedings{Seiler_Pohl_Bossek_Kerschke_Trautmann_2020, place={Berlin, Heidelberg}, title={Deep Learning as a Competitive Feature-Free Approach for Automated Algorithm Selection on the Traveling Salesperson Problem}, DOI={<a href=\"https://doi.org/10.1007/978-3-030-58112-1_4\">10.1007/978-3-030-58112-1_4</a>}, booktitle={Parallel Problem Solving from {Nature} (PPSN XVI)}, publisher={Springer-Verlag}, author={Seiler, Moritz and Pohl, Janina and Bossek, Jakob and Kerschke, Pascal and Trautmann, Heike}, year={2020}, pages={48–64} }"},"publisher":"Springer-Verlag","type":"conference","keyword":["Automated algorithm selection","Deep learning","Feature-based approaches","Traveling Salesperson Problem"],"department":[{"_id":"819"}],"_id":"48897","status":"public","date_created":"2023-11-14T15:59:00Z","doi":"10.1007/978-3-030-58112-1_4","year":"2020","author":[{"last_name":"Seiler","full_name":"Seiler, Moritz","first_name":"Moritz"},{"first_name":"Janina","last_name":"Pohl","full_name":"Pohl, Janina"},{"last_name":"Bossek","full_name":"Bossek, Jakob","first_name":"Jakob","id":"102979","orcid":"0000-0002-4121-4668"},{"last_name":"Kerschke","full_name":"Kerschke, Pascal","first_name":"Pascal"},{"last_name":"Trautmann","full_name":"Trautmann, Heike","first_name":"Heike"}],"publication_identifier":{"isbn":["978-3-030-58111-4"]},"user_id":"102979","title":"Deep Learning as a Competitive Feature-Free Approach for Automated Algorithm Selection on the Traveling Salesperson Problem","publication":"Parallel Problem Solving from {Nature} (PPSN XVI)","extern":"1","abstract":[{"text":"In this work we focus on the well-known Euclidean Traveling Salesperson Problem (TSP) and two highly competitive inexact heuristic TSP solvers, EAX and LKH, in the context of per-instance algorithm selection (AS). We evolve instances with nodes where the solvers show strongly different performance profiles. These instances serve as a basis for an exploratory study on the identification of well-discriminating problem characteristics (features). Our results in a nutshell: we show that even though (1) promising features exist, (2) these are in line with previous results from the literature, and (3) models trained with these features are more accurate than models adopting sophisticated feature selection methods, the advantage is not close to the virtual best solver in terms of penalized average runtime and so is the performance gain over the single best solver. However, we show that a feature-free deep neural network based approach solely based on visual representation of the instances already matches classical AS model results and thus shows huge potential for future studies.","lang":"eng"}],"page":"48–64","language":[{"iso":"eng"}],"place":"Berlin, Heidelberg","date_updated":"2023-12-13T10:49:45Z"}