Dagstuhl Seminar Wiki
- Dagstuhl Seminar Wiki (Use personal credentials as created in DOOR to log in)
- Dagstuhl Materials Page (Use personal credentials as created in DOOR to log in)
The purpose of multiobjective optimization is to develop methods that can solve problems having a number of (conflicting) optimization criteria and constraints, providing a multitude of solution alternatives, rather than pursuing only one "optimal" solution. In this aim the field is highly successful: its methods have a track record of improving decision making across a broad swath of applications, indeed wherever there are conflicting goals or objectives. Yet, since multiobjective optimization has focused almost exclusively on serving a single "decision maker", providing solutions merely as potential (not actual) alternatives, it is not presently a technology that can serve mass markets with mass solutions. A new approach is needed if we are to fulfil the demanding aims of mass-customization, product/service variation and personalization we see today in areas such as engineering, planning, operations, investment, media and Web services, and healthcare. Taking the first steps, this Dagstuhl Seminar will explore an “Analytics” perspective already proven in handling large-scale pervasive data, and seek to build the scientific foundations for delivering efficient and effective (even optimal) mass-personalization.
The seminar will be organized around three application challenges which distinguish between different ways that personalization can be needed or delivered in an optimization and decision-making setting. These are
- platform design and product lines,
- responsive and online personalization, and
- complex networks of decision makers.
All three topics provide a number of challenges in common, as well as distinctive aspects, and all will benefit from the input of researchers from a variety of backgrounds and interests across Evolutionary Multiobjective Optimization (EMO) and Multiple Criteria Decision Making (MCDM).
During this Dagstuhl Seminar, the three application challenges will be crosslinked with three research domains that constitute the methodological core of multiobjective optimization and have been the foundation for the discussions at the previous Dagstuhl Seminars.
- Model building, approximation, and representation comprising theoretical aspects, interwoven systems, set based optimization, interrelating decision space and objective space, robustness and noise handling, and analytics.
- Preference modelling comprising problem structuring and formulation, preference elicitation and learning, preference models and trade-offs, group decision making, high-dimensional problems, interwoven systems, and analytics.
- Algorithm design and efficiency comprising complexity, handling very large data, high-dimensional problems (many objective functions), evolutionary vs exact approaches (and hybrids), and analytics.
This Seminar carries on a series of five previous Dagstuhl Seminars (04461, 06501, 09041, 12041 and 15031) that were focused on Multiobjective Optimization. Our major goal is to further strengthen the links between the EMO and MCDM communities, and to advance both theoretical understanding and computational techniques in multiobjective optimization.
The topic of the seminar, Personalization in Multiobjective Optimization, was motivated by ongoing changes in many areas of human activity. In particular, personalization, mass customization, and mass data have become essential in current business and engineering operations creating new challenges for academic and research communities. In the seminar, the EMO and MCDM communities, including junior and senior academic researchers as well as industry representatives, took an effort to jointly address the ongoing changes in the real-world with multiobjective optimization.
The purpose of multiobjective optimization is to develop methods that can solve problems having a number of (conflicting) optimization criteria and constraints, providing a multitude of solution alternatives, rather than pursuing only one "optimal" solution. In this aim the field has been highly successful: its methods have a track record of improving decision making across a broad swath of applications, indeed wherever there are conflicting goals or objectives. Yet, multiobjective optimization has so far focused almost exclusively on serving a single "decision maker", providing solutions merely as potential (not actual) alternatives. In order to fulfill the demanding aims of mass-customization, product/service variation and personalization we see today in areas such as engineering, planning, operations, investment, media and Web services, and healthcare, new and innovative approaches are needed. This seminar took the first steps towards this goal by bringing together leading specialists in EMO and MCDM.
Personalization in multiobjective optimization as the main theme of the seminar has focused around three application challenges which are highly characteristic for real-world decision making and represent different ways that personalization is needed or delivered in an optimization setting. These were (i) Platform design and product lines, (ii) Responsive and online personalization, and (iii) Complex networks of decision makers. These three application challenges were crosslinked with three research domains that constitute the methodological core of multiobjective optimization and have been the foundation for the discussions at the previous Dagstuhl seminars. These were (1) Model building, (2) Preference modelling, and (3) Algorithm design and efficiency.
During the seminar, we formed five multi-disciplinary working groups (WGs) to implement the crosslinking between these application challenges and research domains, see Table 1. Each working group was focused on an application challenge (a row in Table 1; WGs 2, 3 and 4) or a research domain (a column in Table 1; WGs 1 and 5), all taking specific perspectives on the respective topics.
The program was updated on a daily basis to maintain flexibility in balancing time slots for talks, discussions, and working groups. The working groups were established on the first day in an open and highly interactive discussion. The program included several opportunities to report back from the working groups in order to establish further links and allow for adaptations and feedback. Some of the working groups split into subgroups and rejoined later in order to focus more strongly on different aspects of the topics considered. Abstracts of the talks and extended abstracts of the working groups can be found in subsequent chapters of this report. Further notable events during the week included: (i) a hike on Wednesday afternoon with some sunshine (despite the quite terrible weather during the rest of the week), (ii) an announcements session allowing us to share details of upcoming events in our research community, and (iii) a wine and cheese party made possible by the support of the ITWM Kaiserslautern, represented by Karl-Heinz Küfer.
Fourteen topical presentations were complemented by discussions in five working groups, covering the main themes of the seminar. The outcomes of each of the working groups can be seen in the sequel. Extended versions of their findings will be submitted to a Special Issue on "Personalization in Multiobjective Optimization: An Analytics Perspective" of the Journal of Multicriteria Decision Analysis, edited by Theo Stewart, that is guest edited by the organizers of this seminar. The submission deadline is July 31, 2018, and several working groups plan to submit extended versions of their reports to this special issue.
The seminar was highly productive, very lively and full of discussions, and has thus further strengthened the interaction between the EMO and MCDM communities. We expect that the seminar will initiate a new research domain interrelating multiobjective optimization and personalization, as it similarly has happened after the previous seminars in this series.
A huge thank you to the Dagstuhl office and its very helpful and patient staff; many thanks to the organizers of the previous seminars in the series for the initiative and continuing advice; and many thanks to all the participants, who contributed in so many different ways to make this week a success. In the appendix, we also give special thanks to Joshua Knowles as he steps down from the organizer role.
- Richard Allmendinger (University of Manchester, GB) [dblp]
- Mickaël Binois (Argonne National Laboratory - Lemont, US) [dblp]
- Jürgen Branke (University of Warwick, GB) [dblp]
- Dimo Brockhoff (INRIA Saclay - Palaiseau, FR) [dblp]
- Roberto Calandra (University of California - Berkeley, US) [dblp]
- Carlos A. Coello Coello (CINVESTAV - Mexico, MX) [dblp]
- Kerstin Dächert (Universität Wuppertal, DE) [dblp]
- Kalyanmoy Deb (Michigan State University, US) [dblp]
- Matthias Ehrgott (Lancaster University Management School, GB) [dblp]
- Gabriele Eichfelder (TU Ilmenau, DE) [dblp]
- Michael Emmerich (Leiden University, NL) [dblp]
- Alexander Engau (Lancaster University Management School, GB) [dblp]
- Georges Fadel (Clemson University - Clemson, US) [dblp]
- José Rui Figueira (IST - Lisbon, PT) [dblp]
- Carlos M. Fonseca (University of Coimbra, PT) [dblp]
- Abhinav Gaur (Michigan State University - East Lansing, US) [dblp]
- Salvatore Greco (University of Catania, IT) [dblp]
- Jussi Hakanen (University of Jyväskylä, FI) [dblp]
- Johannes Jahn (Universität Erlangen-Nürnberg, DE) [dblp]
- Andrzej Jaszkiewicz (Poznan University of Technology, PL) [dblp]
- Milosz Kadzinski (Poznan University of Technology, PL) [dblp]
- Kathrin Klamroth (Universität Wuppertal, DE) [dblp]
- Karl Heinz Küfer (Fraunhofer ITWM - Kaiserslautern, DE) [dblp]
- Christoph Lofi (TU Delft, NL) [dblp]
- Manuel López-Ibáñez (University of Manchester, GB) [dblp]
- Kaisa Miettinen (University of Jyväskylä, FI) [dblp]
- Sanaz Mostaghim (Universität Magdeburg, DE) [dblp]
- Boris Naujoks (TH Köln, DE) [dblp]
- Frank Neumann (University of Adelaide, AU) [dblp]
- Luís Paquete (University of Coimbra, PT) [dblp]
- Robin Purshouse (University of Sheffield, GB) [dblp]
- Patrick M. Reed (Cornell University, US) [dblp]
- Günter Rudolph (TU Dortmund, DE) [dblp]
- Stefan Ruzika (TU Kaiserslautern, DE) [dblp]
- Serpil Sayin (Koc University - Istanbul, TR) [dblp]
- Pradyumn Kumar Shukla (KIT - Karlsruher Institut für Technologie, DE) [dblp]
- Roman Slowinski (Poznan University of Technology, PL) [dblp]
- Ralph E. Steuer (University of Georgia, US) [dblp]
- Theodor J. Stewart (University of Cape Town, ZA) [dblp]
- Michael Stiglmayr (Universität Wuppertal, DE) [dblp]
- Lothar Thiele (ETH Zürich, CH) [dblp]
- Selvakumar Ulaganathan (Noesis Solutions - Leuven, BE) [dblp]
- Daniel Vanderpooten (University Paris-Dauphine, FR) [dblp]
- Margaret M. Wiecek (Clemson University, US) [dblp]
- Dagstuhl Seminar 04461: Practical Approaches to Multi-Objective Optimization (2004-11-07 - 2004-11-12) (Details)
- Dagstuhl Seminar 06501: Practical Approaches to Multi-Objective Optimization (2006-12-10 - 2006-12-15) (Details)
- Dagstuhl Seminar 09041: Hybrid and Robust Approaches to Multiobjective Optimization (2009-01-18 - 2009-01-23) (Details)
- Dagstuhl Seminar 12041: Learning in Multiobjective Optimization (2012-01-22 - 2012-01-27) (Details)
- Dagstuhl Seminar 15031: Understanding Complexity in Multiobjective Optimization (2015-01-11 - 2015-01-16) (Details)
- Dagstuhl Seminar 20031: Scalability in Multiobjective Optimization (2020-01-12 - 2020-01-17) (Details)
- Dagstuhl Seminar 23361: Multiobjective Optimization on a Budget (2023-09-03 - 2023-09-08) (Details)
- modelling / simulation
- optimization / scheduling
- soft computing / evolutionary algorithms
- multiobjective optimization
- distributed decision making
- large data