News

HÛ Quentin and Deshayes Baptiste, two students from Nantes Université, are visiting the Multiobjective Optimization Group.

Dr Pavel Bastl and Prof Nirupam Chakraborti will be visiting us from the Czech Technical University to collaborate on the DESIDES project on the 9th and 10th of November.

A new doctoral student, MaoMao Liang, has joined our group. Her research focuses include multiobjective evolutionary algorithms and indicators for interactive methods.

The Finnish Operations Research Society (FORS) organized an anniversary conference to celebrate its 50th anniversary (FORS50). FORS50 occurred at the University of Jyväskylä, Finland, on September 23-24, 2023.

Two of our team members gave talks at this conference:

Babooshka Shavazipour: Decision-making under deep uncertainty: basics, state-of-the-art, and the Future

Juuso Pajasmaa: Challenges of Group Decision-Making in Multiobjective Optimization

For more information, visit: https://www.operaatiotutkimus.fi/index.php/in-english/

We welcome you to visit our group at JYU for the Researchers' Night event. This year, our group is hosting activities related to multiobjective optimization. You can view a video description here.

Assoc. Prof. Hemant Kumar Singh from the University of New South Wales, Australia is visiting the Multiobjective Optimization Group for two months as a University of Jyväskylä Visiting Fellow. He is a leading expert in evolutionary multiobjective optimization.

The paper titled Using a Database to Support Interactive Multiobjective Optimization, Visualization, and Analysis has been presented in the GECCO 2023 conference. The paper is available online on ACM digital library as open access.

Bhupinder Singh Saini, Giomara Lárraga, and Kaisa Miettinen. 2023. Using a Database to Support Interactive Multiobjective Optimization, Visualization, and Analysis. In Proceedings of the Companion Conference on Genetic and Evolutionary Computation (GECCO '23 Companion). Association for Computing Machinery, New York, NY, USA, 1703–1711. https://doi.org/10.1145/3583133.3596383

The two days of the DESDEO 2023 Forum on the future of interactive multiobjective optimization were packed to the brim with interesting talks and presentations. We would like to warmly thank all the invited speakers, presenters, participants, and anybody that helped with the organization of the event. You made our fantastic event possible! Hopefully, this is the first in a series of many more similar events...

Check the event's website for the group photo and the materials section for slides of some of the presentations. Tutorial recordings may be found on our YouTube channel.

Thank you once more!

Follow the forum here: Multiobjective Optimization Group on Twitter.

Today, the Academy of Finland has granted funding for the DESIDES (Decision Support with Interactive Advanced Data-Enabled Multiobjective Optimization Systems) project! This means that DESDEO will be continued to be developed even further with new features and enhancements in the future, including advanced visualizations, explainability tools, and support for group decison-making!

More info about the funding and the funding decisions can be found here.

Adhe Kania from the Multiobjective Optimization Group will defend her PhD thesis Addressing Challenges of Real-World Lot Sizing Problems with Interactive Multiobjective Optimization on Thursday, June 8 at noon in Agora, Aud 2. Live stream is also available here.

DESDEO has played a key role in Adhe's thesis work!

Five speakers have been announced to give a presentation during the DESDEO23 Forum on the future of interactive multiobjective optimization. The speakers and the abstract of their presentation can be found on the event's website here.

Registration to the Forum is still open and we still accepting contributed talks!

Ana Belén Ruiz from the Department of Applied Economics (Mathematics) at the University of Málaga (Spain) is visiting our group this week. Some of Dr. Ruiz's work has utilized DESDEO in the past!

There is a fresh Discord server for discussing topics related to MCDM. It has also a category of channels for discussing DESDEO! If you are having trouble with, or questions about, DESDEO, of just want to hang out with like-minded people, head over to our new server!

Invitation link: https://discord.gg/3baHY5spzC.

Future versions of the DESDEO core packages will support Python versions 3.9 and 3.10. Older versions are no longer officially supported. For individual packages, this means that at least Python version 3.9 is required for the following versions:

• desdeo-tools: >=1.8.0
• desdeo-problem: >=1.5.0
• desdeo-mcdm: >= 1.3.0
• desdeo-emo: >= 1.5.0

From version 1.3.0 onwards, desdeo-mcdm now includes the PAINT method.

Our research group is organizing the Future of Interactive Multiobjective Optimization Forum, which will take place on June 19-20 at the University of Jyväskylä. At this event, you will be able to see demonstrations of the DESDEO framework and learn more about it. More information will be available soon.

We are excited to announce that Theo Hermegil has joined our Multiobjective Optimization Research Group through an Erasmus program. Theo will be working with us for the next 6 months, partly focusing on the development and application of DESDEO. We are thrilled to welcome Theo to our team and look forward to the valuable contributions he will make to our ongoing research efforts in the field of multiobjective optimization!

The Multiobjective Optimization Group is looking for trainees eager to contribute to DESDEO! We offer paid positions for up to three trainees. The trainee-ship will last 3-4 months, but must end by the end of August 2023. See the ad here for more info.

The paper titled Interactive data-driven multiobjective optimization of metallurgical properties of microalloyed steels using the DESDEO framework has recently been published online in EAAI. The paper is available online on ScienceDirect as open access.

Saini, B.S., Chakrabarti, D., Chakraborti, N., Shavazipour, B., Miettinen, K., Interactive data-driven multiobjective optimization of metallurgical properties of microalloyed steels using the DESDEO framework, Engineering Applications of Artificial Intelligence, Volume 120, 105918, 2023.

The DESDEO team will take a break to enjoy the holidays. We will be back next year, in the second week of January. See you then!

Happy holidays from the whole DESDEO team!

This year we received multiple visitors in the multiobjective optimization group at the University of Jyväskyla:

Sunith Bandaru (Associate Professor, University of Skovde, Sweeden) Magda Osman (Professor, University of Cambridge, UK) Ofer M. Shir (Associate Professor, Tel-Hai College, Israel) Henrik Smedberg (Doctoral researcher, Univeristy of Skovde, Sweden) Francisco Ruiz (Professor, University of Malaga, Spain) Luis Paquete (Associate Professor, Univeristy of Coimbra, Portugal)

If you are interested in visiting us, do not hesitate to contact us! :)

Pouya will be defending his thesis Friday on the 2nd of December. The topic of his thesis is Interactive Evolutionary Multiobjective Optimization: Methods and Quality Indicators. The dissertation can be followed online or attended live.

More information can be found here.

The paper titled Designing empirical experiments to compare interactive multiobjective optimization methods has been recently published online in the Journal of the Operational Research Society. The paper is available online as open access.

Afsar, B., Silvennoinen, J., Misitano, G., Ruiz, F., Ruiz, A.B., Miettinen, K., Designing empirical experiments to compare interactive multiobjective optimization methods, Journal of the Operational Research Society, to appear.

Giovanni Misitano was awarded a special grant of 5000€ from the Nokia Foundation for best pitch in IT during the Millennium Graduate Student Contest held during the Millennium Innovation Forum! Giovanni's pitch was about explainable multiobjective optimization. DESDEO has been in a central role in his research developing this new paradigm in decision-support further.

More information about the contest and its results can be found here

For more details about the event, including links to the dissertation and the live broadcast of the public defense, use this link.

We, as our research team, is now back the summer break! During the summer, a couple of papers related to DESDEO have been published.

Misitano, G., Afsar, B., Lárraga, G., and Miettinen, K. Towards explainable interactive multiobjective optimization: R-XIMO. Auton Agent Multi-Agent Syst 36, 43 (2022). link

Mazumdar, A., Otayagich, S., and Miettinen, K. Interactive evolutionary multiobjective optimization with modular physical user interface. GECCO '22: Proceedings of the Genetic and Evolutionary Computation Conference Companion, Pages 1835–1843 (2022). link

Juuso Pajasmaa will be continuing in our research group in the position of research assistant. He will be conducting a literature survey on group decision making in multiobjective optimization.

The paper titled Interactivized: Visual Interaction for Better Decisions With Interactive Multiobjective Optimization has been recently published online in the IEEE Access. The paper is available online IEEE Xplore as open access.

Hakanen, J., Radoš, S., Misitano G., Saini, B. S., Miettinen K., Matković K., Interactivized: Visual Interaction for Better Decisions With Interactive Multiobjective Optimization, IEEE Access, Volume 10, pp. 33661 - 33678, 2022

Our new summer trainee, Matti Viitanen, has started working in our research group. He will work on DESDEO focusing on implementing various test problems. Welcome, Matti!

You can now watch a recording of the DEMO tutorial talk at YouTube and access the slides at GitHub.

The members of the Multiobjective Optimization Group, including members of the DESDEO team, will teach the "Interactive Multiobjective Optimization: Applications and Tools to Support Decision Making" course in the 31st Jyväskylä Summer School. Find more information here.

JYU published a press release about a joint research study conducted in our group collaborating with the Faculty of Sports Science scientists.

For further information, see the press release here.

We are looking for an eager student (currently enrolled at JYU) to work in our research group as a trainee during the summer. The trainee will be contributing to DESDEO by working with various aspects related to developing and testing the framework.

For further information, see the ad here.

The paper titled Interactive multiobjective optimization for finding the most preferred exercise therapy modality in knee osteoarthritis has been published online in the journal called Annals of Medicine. The paper is available online via Taylor & Francis (open access).

Shavazipour, B., Afsar, B., Multanen, J., Miettinen, K., Kujala, U.M., Interactive multiobjective optimization for finding the most preferred exercise therapy modality in knee osteoarthritis, Annals of Medicine, 54(1), 181-194, 2022, DOI.

The paper titled Optimistic NAUTILUS Navigator for Multiobjective Optimization with Costly Function Evaluations has been published online in the Journal of Global Optimization. The paper is available online via SpringerLink (open access).

Saini, B. S., Emmerich, M., Mazumdar, A., Afsar, B., Shavazipour, B., Miettinen, K., Optimistic NAUTILUS Navigator for Multiobjective Optimization with Costly Function Evaluations, Journal of Global Optimization, 2022.

The DESDEO team will be taking their well deserved rest and enjoy the holidays. We will be back next year on the second week of January. See you then!

Happy holidays from the whole team!

The paper titled Comparing interactive evolutionary multiobjective optimization methods with an artificial decision maker has been published online in the journal called Complex & Intelligent Systems. The paper is available online via SpringerLink.

Afsar, B., Ruiz, A.B., Miettinen, K., Comparing interactive evolutionary multiobjective optimization methods with an artificial decision maker, Complex & Intelligent Systems, 2021, DOI.

Tuomo Kalliokoski gave a tutorial on eXplainable Artificial Intelligence during our weekly DEMO seminar. The recording of the tutorial is available on YouTube.

Giovanni Misitano gave a talk during the DEMO weekly seminars about DESDEO. The tutorial was very practice oriented with a lot of code shown. The material covered defining multiobjective optimization problems, solving them by scalarization and (very basic) evolutionary methods, and how to interact with interactive multiobjective optimization methods.

The talk was recorded. The recording is available on YouTube.

The slides of the presentation are available here.

The Python notebook shown in the presentation is available on GitHub.

This week, Giovanni Misitano presented a poster titled Explainable Interactive Multiobjective Optimization during the IT Faculty Research Day 2021 held at the University of Jyväskylä on November 24th. The poster presents a new exciting idea of applying SHAP values in interactive multiobjective optimization to explain reference point based methods. The work presented in the poster is related to an active research topic involving also Bekir Afsar, Giomara Lárraga Maldonado, and Kaisa Miettinen.

If you missed the poster presentation, the poster is available online here.

The DEMO Tutorial given by Dr. Babooshka Shavazipour on “Decision Making under Deep Uncertainty” can be viewed using the following link: https://youtu.be/OiGVwncLBi0.

Our article title DESDEO: The Modular and Open Source Framework for Interactive Multiobjective Optimization has been published in IEEE Access! Check it out and remember to cite it if you happen to use DESDEO in your work!

The article is available as open-acces here.

We have updated the DESDEO website with some changes and new features. These include:

• A new section of the website: Materials. Here, you can find various files and other assets related to DESDEO that are shared publicly.
• News articles can now be linked to directly! Just copy and paste the url found in the link after the date in the news article heading.
• You can now navigate to different years in the news.
• Figures can now be attached to new. So expect more visual content to be found in the news from now on!
• Navigation utilities added to the publications sections as well. You can now jump to a specific publication type.

Here is a figure to showcase the new possibility to have figures in the news:

New MSc thesis topics are available! The contents of the topics are related to the DEMO thematic research area! Many, if not most, of the topics are very closely related to DESDEO. So! If you are looking to complete your master's degree and to contribute to DESDEO, look no further! The topics are suggestive, which means that they can be tailored to meet your goals and interests. For example, you can choose to focus on the software side or on the theoretical side in some of the topics.

For the list of the available topics, see this document!

Giovanni Misitano gave a talk about his personal experiences in developing DESDEO with a particular focus on the different interfaces that have been developed for DESDEO in the past. Old iterations of DESDEO are also discussed.

The talk was recorded and is available on YouTube.

The paper titled Comparing Interactive Evolutionary Multiobjective Optimization Methods with an Artificial Decision Maker has been accepted in the journal called Complex & Intelligent Systems.

Afsar, B., Ruiz, A.B., Miettinen, K., Comparing Interactive Evolutionary Multiobjective Optimization Methods with an Artificial Decision Maker, Complex & Intelligent Systems, to appear, 2021.

This week has been very eventful regarding recent developments in DESDEO's graphical interfaces. First, a tutorial was given by Giovanni Misitano about the web stack recently developed to interface into DESDEO, and that can be used to build graphical interfaces for DESDEO. The three components of the stack are available on GitHub: desdeo-webapi, desdeo-components, and desdeo-frontend. The tutorial given by Giovanni was recorded and is available on YouTube.

Second, our summer trainee Juuso Pajasmaa gave a seminar talk about his experiences during the summer in our weekly DEMO seminar slot. Juuso worked on implementing new visualizations in desdeo-components and also implemented a working interface for NAUTILUS Navigator in desdeo-frontend. Moreover, he made contribution to DESDEO working on implementing indicators for PBEA/IBEA and implementing a DBMOPP generator, among other things as well. Juuso's talk is available on YouTube. Juuso will continue working on DESDEO also in the future!

The paper titled Interactive Multiobjective Optimization in Lot Sizing with Safety Stock and Safety Lead Time. has been published in the Proceedings of the In International Conference on Computational Logistics (ICCL 2021). The paper can be found online here.

Kania A., Sipilä J., Afsar B., Miettinen K. (2021) Interactive Multiobjective Optimization in Lot Sizing with Safety Stock and Safety Lead Time. In: Mes M., Lalla-Ruiz E., Voß S. (eds) Computational Logistics (pp. 208-221). ICCL 2021. Lecture Notes in Computer Science, vol 13004. Springer, Cham.

During this autumn, Jana Burkotová, assistant professor at the Faculty of Science of Palacký University, Czech Republic, has been visiting the Multiobjective Optimization Group. Her research interest concerns numerical methods and optimization. Currently, she is turning her specialization towards multiobjective optimization problems. In particular, she focuses on computationally expensive optimization problems with applications in the pump industry. Last Wednesday, she gave a talk in our DEMO seminar about her research interest in the field of pump industry

Michael Emmerich is an associate professor at Leiden University (The Netherlands) currently visiting the Multiobjective Optimization Group as a research fellow. He gave a talk in our weekly seminars titled Towards Explainable Artificial Intelligence in Multiobjective Optimization. In his talk, Michael Emmerich gave a brief introduction about explainable artificial intelligence and gave some exciting ideas on how to apply it in multiobjective optimization.

The talk is available for viewing on YouTube.

One of the trainees working in our research groups during the summer, Stefan Otayagich, has developed modular physical interfaces to control preference input in interactive multiobjective optimization methods. The layout of the interface can be modified by adding, removing, or interchanging physical modules containing switches, nobs, buttons and sliders. As far as we know, Stefan's work is first of its kind!

Last Wednesday, Stefan gave a talk about his work during the summer in our weekly DEMO seminar. The whole talk was recorded and is available to be viewed on YouTube. The code related to this project can be found on GitHub.

Sadly, with this talk Stefan's journey within our research group came to its end (at least for now). On behalf of our research group, I would like to thank Stefan for his massive efforts and great work. He managed to overcome all expectations!

Thank you Stefan!

The 30th Jyväskylä Summer School ended on the 20th of August. Students enrolled in the Multicriteria Design Optimization course (lectured by Assoc. Prof. Michael Emmerich, Leiden University) successfully used the DESDEO framework to learn about multiobjective optimization methods and solve Design Optimization problems. The codes and slides can be found on GitHub.

Our research group at the University of Jyväskylä is back from the summer holidays. During the summer, a conference paper title Interactive multiobjective optimization in lot sizing with safety stock and safety lead time by Adhe Kania, Juha Sipilä, Bekir Afsar, and Kaisa Miettinen was accepted to be presented at the International Conference on Computational Logistics in September. The paper will be published in the conference proceedings published by Springer. DESDEO has been utilized in this work.

The paper titled Multi-scenario multi-objective robust optimization under deep uncertainty: A posteriori approach has been published online in the journal called Environmental Modelling & Software. The paper is openly available via sciencedirect.

B. Shavazipour, J. H. Kwakkel, K. Miettinen, Multi-scenario multi-objective robust optimization under deep uncertainty: A posteriori approach, Environmental Modelling & Software, Volume 144, 2021, 105134, DOI.

The paper titled Visualizations for decision support in scenario-based multiobjective optimization has been published online in the journal called Information Sciences. The paper is openly available via sciencedirect.

B. Shavazipour, M. López-Ibáñez, K. Miettinen, Visualizations for decision support in scenario-based multiobjective optimization, Information Sciences, Volume 578, Pages 1-21, 2021. DOI.

Two new trainees have joined us for the summer and will work on DESDEO. Juuso Pajasmaa is working on graphical interfaces and new interactive visualizations implemented in Typescript, React, and D3. While Stefan Otayagich is working on physical interfaces to be used with interactive multiobjective optimization methods found in DESDEO using Arduinos. Both of our trainees will also work on the core packages in DESDEO implementing both interactive MCDM and EMO methods not yet present in DESDEO. It is safe to say we all expect exciting advancements in the development of DESDEO to take place during this summer!

We have updated the documentation of DESDEO framework. Have a look at the latest version here.

Our letter of intent sent in the first stage for applying as a Centre of Excellence in a funding call by the Academy of Finland was accepted. This means our research group has now a shot at being nominated as a Finnish Centre of Excellence for the period 2022 - 2029. Being chosen as a Centre of Excellence is one of the biggest merits a Finnish research group can earn. Fingers crossed!

Read more about Centres of Excellence here.

The paper titled Assessing the Performance of Interactive Multiobjective Optimization Methods: A Survey has been published online in the journal called ACM Computing Surveys. The paper is available via ACM Digital Library and JYX.

Afsar, B., Miettinen, K., Ruiz, F., Assessing the Performance of Interactive Multiobjective Optimization Methods: A Survey, ACM Computing Surveys, 54(4), article 85, 2021.

Two trainees from Iran joined our group a few weeks back to attend a trainee-ship. Due to the pandemic, they have been working remotely on problem formulation and method implementation in DESDEO. A brief introduction of the two trainees is given below:

Golman Rahmanifar is currently studying transport system engineer at the Sapienza University of Rome. Her expertise lies mainly in operations research (more precisely in multi-objective optimization) and its application to transportation, scheduling, supply chain management, operation management, etc. She is interested in solving challenging real-life optimization problems, specifically in the logistics of freight transportation, in order to help solving the world's challenges, investigate behavioral aspects of supply chain decision making, and develop mathematical models to forecast flows or optimize logistic networks.

Mostafa Mohammadi is also currently studying transportation system engineering at the Sapienza University of Rome. He has eight years of experience in manufacturing and distribution companies, and is interested in applying multi-objective optimization and its application in supply chain management, transportation, and scheduling to optimize real-world problems in industry.

The DEMO Tutorial gave by Dr. Jussi Hakanen on “Tutorial on Visualization for Decision Support in Many-Objective Optimization” can be viewed using the following link: https://youtu.be/NTGL4BpgpKI.

The new release brings support for a family of scalarization functions using the parameterized GLIDE-II framework. Fast non-dominated sorting has also been implemented, removing the requirement for the external PyGMO package.

The paper titled An Artificial Decision Maker for Comparing Reference Point Based Interactive Evolutionary Multiobjective Optimization Methods has been published in the Proceedings of the 11th International Conference on Evolutionary Multi-Criterion Optimization (EMO 2021). The paper can be found online here.

UPDATE 26.3.2021: Thank you for your interest! The call has ended.

The Multiobjective Optimization Group is looking for a summer trainee interested in developing web-interfaces for using interactive methods found in the DESDEO framework. Read more about it here.

The DEMO Tutorial gave by Prof. Kaisa Miettinen on "(Interactive) Multiobjective Optimization" can be viewed using the following link: https://youtu.be/JK_jJBK5Zgc.

The latest release of desdeo-emo brings support for MOEA/D and associated scalarizing functions! The scalarizing functions are implemented in a modular fashion and can be found in the desdeo-tools package. Read more details in the documentation. Comments and suggestions are welcome!

The new release brings support for new methods of preference elicitation such as choosing preferred or non-preferred solutions. Read more details in the documentation. Comments and suggestions welcome!

The DEMO Seminar talk given by Giovanni Misitano and Bhupinder Saini on "How to develop and contribute to DESDEO" can be viewed using the following link: https://youtu.be/JCZRBx55KBQ.

The paper titled Assessing the Performance of Interactive Multiobjective Optimization Methods: A Survey has been accepted in the journal called ACM Computing Surveys.

Afsar, B., Miettinen, K., Ruiz, F., Assessing the Performance of Interactive Multiobjective Optimization Methods: A Survey, ACM Computing Surveys, to appear, 2021.

The paper titled Surrogate assisted interactive multiobjective optimization in energy system design of buildings has been published online in the journal called Optimization and Engineering. The paper is openly accessible via SpringerLink.

Aghaei Pour, P., Rodemann, T., Hakanen, J., Miettinen, K. Surrogate assisted interactive multiobjective optimization in energy system design of buildings. Optim Eng (2021).

The paper titled Interactively Learning the Preferences of a Decision Maker in Multi-objective Optimization Utilizing Belief-rules by Giovanni Misitano has been published in the Proceedings of the 2020 IEEE Symposium Series on Computational Intelligence (SSCI). The paper can be found online here.

We are delighted to announce the acceptance of our paper for the EMO2021 conference.

Afsar, B., Miettinen, K., Ruiz, A. B., Artificial Decision Maker for Comparing Reference Point Based Interactive Evolutionary Multiobjective Optimization Methods, in "Evolutionary Multi-Criterion Optimization, 11th International Conference, Proceedings", Springer, to appear, 2021.

The desdeo-tools package now contains support for the epsilon constraint scalarization method. The desdeo-mcdm package has now an implementation of NAUTILUS ready to be used. These features were contributed by Antti Luopajärvi.

DESDEO v1.1 adds support for popular evolutionary algorithms such as RVEA and NSGA-III. Data-driven optimization and constrained optimization are also supported when using the EAs. Read the documentation for more information. Comments and contributions are welcome!

The first release version of DESDEO, version 1.0, has been released. See the "Software" section of this webpage for further details. We warmly invite anyone interested to try out the framework, give their feedback, and contribute to the project! There is still a lot to do, and many details to be ironed out. This is only the beginning of the project!

NAUTILUS Navigator has been implemented in DESDEO. A web-based UI can also be found online to test the method. You can find the demo here: https://dash.misitano.xyz/

The development of interactive multi-objective optimization methods for the desdeo-mcdm module has begun. Upcoming methods include: Synchronous NIMBUS, NAUTILUS, and E-NAUTILUS. The methods will be available by the end of April.

The new DESDEO website has been launched on the World Wide Web.