A4 Artikkeli konferenssijulkaisussa
Incorporating Preference Information Interactively in NSGA-III by the Adaptation of Reference Vectors (2023)
Lárraga, G., Saini, B. S., & Miettinen, K. (2023). Incorporating Preference Information Interactively in NSGA-III by the Adaptation of Reference Vectors. In M. Emmerich, A. Deutz, H. Wang, A. V. Kononova, B. Naujoks, K. Li, K. Miettinen, & I. Yevseyeva (Eds.), Evolutionary Multi-Criterion Optimization : 12th International Conference, EMO 2023, Leiden, The Netherlands, March 20–24, 2023, Proceedings (pp. 578-592). Springer. Lecture Notes in Computer Science, 13970. https://doi.org/10.1007/978-3-031-27250-9_41
JYU-tekijät tai -toimittajat
Julkaisun tiedot
Julkaisun kaikki tekijät tai toimittajat: Lárraga, Giomara; Saini, Bhupinder Singh; Miettinen, Kaisa
Emojulkaisu: Evolutionary Multi-Criterion Optimization : 12th International Conference, EMO 2023, Leiden, The Netherlands, March 20–24, 2023, Proceedings
Emojulkaisun toimittajat: Emmerich, Michael; Deutz, André; Wang, Hao; Kononova, Anna V.; Naujoks, Boris; Li, Ke; Miettinen, Kaisa; Yevseyeva, Iryna
Konferenssin paikka ja aika: Leiden, The Netherlands, 20.-24.3.2023
ISBN: 978-3-031-27249-3
eISBN: 978-3-031-27250-9
Lehti tai sarja: Lecture Notes in Computer Science
ISSN: 0302-9743
eISSN: 1611-3349
Julkaisuvuosi: 2023
Ilmestymispäivä: 21.02.2023
Sarjan numero: 13970
Artikkelin sivunumerot: 578-592
Kirjan kokonaissivumäärä: 636
Kustantaja: Springer
Kustannuspaikka: Cham
Julkaisumaa: Sveitsi
Julkaisun kieli: englanti
DOI: https://doi.org/10.1007/978-3-031-27250-9_41
Julkaisun avoin saatavuus: Ei avoin
Julkaisukanavan avoin saatavuus:
Julkaisu on rinnakkaistallennettu (JYX): https://jyx.jyu.fi/handle/123456789/87520
Tiivistelmä
Real-world multiobjective optimization problems involve decision makers interested in a subset of solutions that meet their preferences. Decomposition-based multiobjective evolutionary algorithms (or MOEAs) have gained the research community’s attention because of their good performance in problems with many objectives. Some efforts have been made to propose variants of these methods that incorporate the decision maker’s preferences, directing the search toward regions of interest. Typically, such variants adapt the reference vectors according to the decision maker’s preferences. However, most of them can consider a single type of preference, the most common being reference points. Interactive MOEAs aim to let decision-makers provide preference information progressively, allowing them to learn about the trade-offs between objectives in each iteration. In such methods, decision makers can provide preferences in multiple ways, and it is desirable to allow them to select the type of preference for each iteration according to their knowledge. This article compares three interactive versions of NSGA-III utilizing multiple types of preferences. The first version incorporates a mechanism that adapts the reference vectors differently according to the type of preferences. The other two versions convert the preferences from the type selected by the decision maker to reference points, which are then utilized in two different reference vector adaptation techniques that have been used in a priori MOEAs. According to the results, we identify the advantages and drawbacks of the compared methods.
YSO-asiasanat: monitavoiteoptimointi; päätöksenteko; päätöksentukijärjestelmät; interaktiivisuus; evoluutiolaskenta
Vapaat asiasanat: multiobjective optimization; interactive methods; decision making; multiobjective evolutionary algorithms; decomposition-based MOEAs; NSGA-III
Liittyvät organisaatiot
Hankkeet, joissa julkaisu on tehty
- Datapohjainen päätöksenteon tuki monitavoiteoptimoinnin avulla (DAEMON)
- Miettinen, Kaisa
- Suomen Akatemia
OKM-raportointi: Kyllä
Raportointivuosi: 2023
Alustava JUFO-taso: 1
Emojulkaisu, jossa JYU-tekijöitä:
- Emmerich, M., Deutz, A., Wang, H., Kononova, A. V., Naujoks, B., Li, K., Miettinen, K., & Yevseyeva, I. (Eds.). (2023). Evolutionary Multi-Criterion Optimization : 12th International Conference, EMO 2023, Leiden, The Netherlands, March 20–24, 2023, Proceedings. Springer. Lecture Notes in Computer Science, 13970. https://doi.org/10.1007/978-3-031-27250-9