Dagstuhl Seminar 27291
Theory of Randomized Optimization Heuristics
( Jul 18 – Jul 23, 2027 )
Permalink
Organizers
- Youhei Akimoto (University of Tsukuba, JP)
- Tobias Glasmachers (Ruhr-Universität Bochum, DE)
- Martin S. Krejca (Ecole Polytechnique - Palaiseau, FR)
- Christine Zarges (Aberystwyth University, GB)
Contact
- Michael Gerke (for scientific matters)
- Jutka Gasiorowski (for administrative matters)
Randomized optimization heuristics (ROHs) are general-purpose optimizers that are widely and successfully applied to both discrete and continuous problems. Although ROHs are not created with a theoretical analysis in mind but are instead built for efficiency on real-world problems, the theoretical study of ROHs has contributed substantially for more than two decades to our understanding of the benefits and shortcomings of ROHs.
This Dagstuhl Seminar marks the latest entry in a sequence of very successful Dagstuhl Seminars dedicated to the theoretical analysis of ROHs. In this seminar, we will focus on three topics in this domain that have also received increased interest over the last years and were not featured in this level of detail in any of the previous seminars of this series: (1) multi-objective optimization, (2) mixed-variable optimization, and (3) optimization under uncertainty.
Multi-objective optimization considers problems with multiple, conflicting objectives. Consequently, solutions are no longer necessarily comparable with each other, and the aim is to find optimal tradeoffs in the objectives, known as the Pareto optima. The theoretical analysis of ROHs has seen a drastic surge within the last few years, but there are still many questions unanswered, even on fundamental aspects. Mixed-variable optimization considers problems that feature both continuous and discrete variables. ROHs for mixed problems recently attracted considerable attention from the community. This seminar thus offers the perfect opportunity to bring together experts from both domains, and we believe that it will foster the recently more strongly emerging collaboration between both domains.
Optimization under uncertainty considers problems where the data or its evaluation is subject to missing information, such as in the case of noise. Since this is a topic that also shows up in many other areas of science, there exists a plethora of approaches to deal with uncertainty, many of which have not been considered with theoretical rigor yet.
As with each iteration, we are looking forward to having a lively exchange between the diverse set of participants. We are very happy to welcome experts from within the ROH community, experts from neighboring areas, as well as junior researchers. There is plenty of room to have engaging discussions on various interesting and important topics that can lead to new collaborations and inspiring ideas.

Related Seminars
- Dagstuhl Seminar 00071: Theory of Evolutionary Algorithms (2000-02-13 - 2000-02-18) (Details)
- Dagstuhl Seminar 02031: Theory of Evolutionary Algorithms (2002-01-13 - 2002-01-18) (Details)
- Dagstuhl Seminar 04081: Theory of Evolutionary Algorithms (2004-02-15 - 2004-02-20) (Details)
- Dagstuhl Seminar 06061: Theory of Evolutionary Algorithms (2006-02-05 - 2006-02-10) (Details)
- Dagstuhl Seminar 08051: Theory of Evolutionary Algorithms (2008-01-27 - 2008-02-01) (Details)
- Dagstuhl Seminar 10361: Theory of Evolutionary Algorithms (2010-09-05 - 2010-09-10) (Details)
- Dagstuhl Seminar 13271: Theory of Evolutionary Algorithms (2013-06-30 - 2013-07-05) (Details)
- Dagstuhl Seminar 15211: Theory of Evolutionary Algorithms (2015-05-17 - 2015-05-22) (Details)
- Dagstuhl Seminar 17191: Theory of Randomized Optimization Heuristics (2017-05-07 - 2017-05-12) (Details)
- Dagstuhl Seminar 19431: Theory of Randomized Optimization Heuristics (2019-10-20 - 2019-10-25) (Details)
- Dagstuhl Seminar 22081: Theory of Randomized Optimization Heuristics (2022-02-20 - 2022-02-25) (Details)
- Dagstuhl Seminar 24271: Theory of Randomized Optimization Heuristics (2024-06-30 - 2024-07-05) (Details)
Classification
- Neural and Evolutionary Computing
Keywords
- black-box optimization heuristics
- evolution strategies
- genetic and evolutionary algorithms
- runtime and convergence analysis
- stochastic processes