Standard

Ordered Preference Elicitation Strategies for Supporting Multi-Objective Decision Making. / Zintgraf, Luisa; Roijers, Diederik; Linders, Sjoerd; Jonker, Catholijn M.; Nowe, Ann.

AAMAS 2018: Proceedings of the Seventeenth International Joint Conference on Autonomous Agents and Multi-Agent Systems. 2018.

Research output: Chapter in Book/Report/Conference proceedingConference paperResearch

Harvard

Zintgraf, L, Roijers, D, Linders, S, Jonker, CM & Nowe, A 2018, Ordered Preference Elicitation Strategies for Supporting Multi-Objective Decision Making. in AAMAS 2018: Proceedings of the Seventeenth International Joint Conference on Autonomous Agents and Multi-Agent Systems. 17th International Conference on Autonomous Agents and Multiagent Systems: AAMAS 2018, Stockholm, Sweden, 10/07/18.

APA

Zintgraf, L., Roijers, D., Linders, S., Jonker, C. M., & Nowe, A. (2018). Ordered Preference Elicitation Strategies for Supporting Multi-Objective Decision Making. In AAMAS 2018: Proceedings of the Seventeenth International Joint Conference on Autonomous Agents and Multi-Agent Systems

Vancouver

Zintgraf L, Roijers D, Linders S, Jonker CM, Nowe A. Ordered Preference Elicitation Strategies for Supporting Multi-Objective Decision Making. In AAMAS 2018: Proceedings of the Seventeenth International Joint Conference on Autonomous Agents and Multi-Agent Systems. 2018

Author

Zintgraf, Luisa ; Roijers, Diederik ; Linders, Sjoerd ; Jonker, Catholijn M. ; Nowe, Ann. / Ordered Preference Elicitation Strategies for Supporting Multi-Objective Decision Making. AAMAS 2018: Proceedings of the Seventeenth International Joint Conference on Autonomous Agents and Multi-Agent Systems. 2018.

BibTeX

@inproceedings{55cdea5968dc41c997bb5427775aa9d2,
title = "Ordered Preference Elicitation Strategies for Supporting Multi-Objective Decision Making",
abstract = "In multi-objective decision planning and learning, much attentionis paid to producing optimal solution sets that contain an optimalpolicy for every possible user preference profile. We argue thatthe step that follows, i.e, determining which policy to execute bymaximising the user’s intrinsic utility function over this (possiblyinfinite) set, is under-studied. This paper aims to fill this gap. Webuild on previous work on Gaussian processes and pairwise comparisons for preference modelling, extend it to the multi-objectivedecision support scenario, and propose new ordered preferenceelicitation strategies based on ranking and clustering. Our maincontribution is an in-depth evaluation of these strategies using computer and human-based experiments. We show that our proposedelicitation strategies outperform the currently used pairwise meth-ods, and found that users prefer ranking most. Our experimentsfurther show that utilising monotonicity information in GPs byusing a linear prior mean at the start and virtual comparisons tothe nadir and ideal points, increases performance. We demonstrateour decision support framework in a real-world study on trafficregulation, conducted with the city of Amsterdam",
author = "Luisa Zintgraf and Diederik Roijers and Sjoerd Linders and Jonker, {Catholijn M.} and Ann Nowe",
year = "2018",
month = "7",
language = "English",
isbn = "978-1-4503-5649-7",
booktitle = "AAMAS 2018: Proceedings of the Seventeenth International Joint Conference on Autonomous Agents and Multi-Agent Systems",

}

RIS

TY - GEN

T1 - Ordered Preference Elicitation Strategies for Supporting Multi-Objective Decision Making

AU - Zintgraf, Luisa

AU - Roijers, Diederik

AU - Linders, Sjoerd

AU - Jonker, Catholijn M.

AU - Nowe, Ann

PY - 2018/7

Y1 - 2018/7

N2 - In multi-objective decision planning and learning, much attentionis paid to producing optimal solution sets that contain an optimalpolicy for every possible user preference profile. We argue thatthe step that follows, i.e, determining which policy to execute bymaximising the user’s intrinsic utility function over this (possiblyinfinite) set, is under-studied. This paper aims to fill this gap. Webuild on previous work on Gaussian processes and pairwise comparisons for preference modelling, extend it to the multi-objectivedecision support scenario, and propose new ordered preferenceelicitation strategies based on ranking and clustering. Our maincontribution is an in-depth evaluation of these strategies using computer and human-based experiments. We show that our proposedelicitation strategies outperform the currently used pairwise meth-ods, and found that users prefer ranking most. Our experimentsfurther show that utilising monotonicity information in GPs byusing a linear prior mean at the start and virtual comparisons tothe nadir and ideal points, increases performance. We demonstrateour decision support framework in a real-world study on trafficregulation, conducted with the city of Amsterdam

AB - In multi-objective decision planning and learning, much attentionis paid to producing optimal solution sets that contain an optimalpolicy for every possible user preference profile. We argue thatthe step that follows, i.e, determining which policy to execute bymaximising the user’s intrinsic utility function over this (possiblyinfinite) set, is under-studied. This paper aims to fill this gap. Webuild on previous work on Gaussian processes and pairwise comparisons for preference modelling, extend it to the multi-objectivedecision support scenario, and propose new ordered preferenceelicitation strategies based on ranking and clustering. Our maincontribution is an in-depth evaluation of these strategies using computer and human-based experiments. We show that our proposedelicitation strategies outperform the currently used pairwise meth-ods, and found that users prefer ranking most. Our experimentsfurther show that utilising monotonicity information in GPs byusing a linear prior mean at the start and virtual comparisons tothe nadir and ideal points, increases performance. We demonstrateour decision support framework in a real-world study on trafficregulation, conducted with the city of Amsterdam

M3 - Conference paper

SN - 978-1-4503-5649-7

BT - AAMAS 2018: Proceedings of the Seventeenth International Joint Conference on Autonomous Agents and Multi-Agent Systems

ER -

ID: 36362723