Working Group 5
Algorithm Selection and Configuration
Leader: Carola Doerr, FR
Vice-Leader: Pascal Kerschke, DE
WG5 focuses on selecting and configuring appropriate Randomised Optimisation Algorithms (ROAs) for given problem instances based on instance features, configuring ROAs that achieve a good level of performance on broad families of instances, and exploring the corresponding trade-offs.
Tasks
- Characterising optimisation problems by analysing the problem landscape (e.g., the traditional fitness landscape in the case of single-objective problems, multiobjective landscape for problems with multiple objectives, violation landscape for constrained problems) and extracting problem features that are useful for better problem understanding.
- Building prediction models that relate the problem features and the performance of various ROAs. These models will then serve to support the selection of appropriate ROAs for newly encountered optimisation problems based on their features.
- Automating the configuration of ROAs, i.e., determining the types of their components, such as the initialisation procedure and operators, as well as the values of the algorithm parameters. For this purpose, various approaches will be studied, ranging from deploying local search techniques to building complex performance models, as well as combining algorithm configuration with algorithm selection.
News
2025-03-27
The development repository for the ROAR-NET API Specification is now publicly available on GitHub.
2025-01-09
ROAR-NET invites applications for Young Researcher and Innovator Conference Grants to be submitted at any time until 30 June 2025.
2024-12-23
ROAR-NET invites applications for Short-Term Scientific Missions to be submitted at any time until 15 July 2025.
Publications
Andova, A., Cork, J. N., Tušar, T., & Filipič, B. (2024). Enhancing algorithm performance prediction in constrained multiobjective optimization using additional training problems. GECCO ’24: Proceedings of the Genetic and Evolutionary Computation Conference, 458–466. https://doi.org/10.1145/3638529.3654098Bibtex
@inproceedings{Andova2024Enhancing,
address = {Melbourne, Australia},
author = {Andova, Andrejaana and Cork, Jordan N. and Tu{\v s}ar, Tea and Filipi{\v c}, Bogdan},
booktitle = {GECCO \textquoteright{}24: Proceedings of the genetic and evolutionary computation conference},
doi = {10.1145/3638529.3654098},
year = {2024},
month = {7},
pages = {458--466},
organization = {ACM},
title = {Enhancing algorithm performance prediction in constrained multiobjective optimization using additional training problems},
}
Djukanović, M., Kartelj, A., Eftimov, T., Reixach, J., & Blum, C. (2024). Efficient search algorithms for the restricted longest common subsequence problem. In L. Franco, C. de Mulatier, M. Paszynski, V. V. Krzhizhanovskaya, J. J. Dongarra, & P. M. A. Sloot (Eds.), Computational Science – ICCS 2024 (Vol. 14836, pp. 58–73). Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-63775-9_5Bibtex
@inbook{Djukanovic2024Efficient,
address = {Cham},
author = {Djukanovi{\' c}, Marko and Kartelj, Aleksandar and Eftimov, Tome and Reixach, Jaume and Blum, Christian},
series = {Lecture {Notes} in {Computer} {Science}},
booktitle = {Computational {Science} -- {ICCS} 2024},
doi = {10.1007/978-3-031-63775-9_5},
editor = {Franco, Leonardo and Mulatier, Cl{\' e}lia and Paszynski, Maciej and Krzhizhanovskaya, Valeria V. and Dongarra, Jack J. and Sloot, Peter M. A.},
isbn = {978-3-031-63775-9},
year = {2024},
month = {7},
pages = {58--73},
publisher = {Springer Nature Switzerland},
title = {Efficient search algorithms for the restricted longest common subsequence problem},
volume = {14836},
}
Pluhacek, M., Viktorin, A., & Senkerik, R. (2024). Designing metaheuristics with large language models: Challenges and opportunities. The IEEE World Congress on Computational Intelligence (IEEE WCCI 2024), 105.Bibtex
@inproceedings{Pluhacek2024Designing,
address = {Yokohama, Japan},
author = {Pluhacek, Michal and Viktorin, Adam and Senkerik, Roman},
booktitle = {The {IEEE} world congress on computational intelligence ({IEEE} {WCCI} 2024)},
year = {2024},
month = {6},
note = {Tutorial},
pages = {105},
title = {Designing metaheuristics with large language models: Challenges and opportunities},
}
Ulutas, E. (2024). A variety of solutions of the nonlinear Schrödinger equation in a white noise environment. The 3rd International Conference on Applied Mathematics in Engineering (ICAME’24), 52.Bibtex
@inproceedings{Ulutas2024variety,
address = {Bal\i{}kesir, T{\" u}rkiye},
author = {Ulutas, Esma},
booktitle = {The 3rd international conference on applied mathematics in engineering ({ICAME}\textquoteright{}24)},
year = {2024},
month = {6},
pages = {52},
title = {A variety of solutions of the nonlinear {Schr}{\" o}dinger equation in a white noise environment},
}
