Planning with preferences using Multi-Attribute Utility Theory along with a Choquet Integral - Archive ouverte HAL Access content directly
Journal Articles Engineering Applications of Artificial Intelligence Year : 2019

Planning with preferences using Multi-Attribute Utility Theory along with a Choquet Integral

(1, 2) , (1) , (2)
1
2

Abstract

This paper addresses the problem of planning with preferences using Multiple Criteria Decision Analysis (mcda) mechanisms. We start by explaining how pddl3 preferences can be modelled by criteria from the Multi-Attribute Utility Theory (maut) along with a Choquet integral. Interestingly, preferences formalized using maut have almost the same expressiveness as the ones formalized in pddl3 while being much easier to model. Next, we present a new heuristic for planning with preferences which is based on the Choquet integral. Finally, we introduce ChoPlan a proof-of-concept planner solving maut-encoded planning problems using the aforementioned heuristic. ChoPlan’s performances are evaluated with respect to state of the art planners using problems from the fifth International Planning Competition.
Fichier principal
Vignette du fichier
planning-with-preferences-using-multi-attribute-utility.pdf (489.82 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-02271098 , version 1 (22-11-2019)

Identifiers

Cite

Loïc Bidoux, Jean-Paul Pignon, Frederick Benaben. Planning with preferences using Multi-Attribute Utility Theory along with a Choquet Integral. Engineering Applications of Artificial Intelligence, 2019, 85, pp.808-817. ⟨10.1016/j.engappai.2019.08.002⟩. ⟨hal-02271098⟩
65 View
82 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More