Refining HTN Methods via Task Insertion with Preferences

Zhanhao Xiao (1, 2) , Hai Wan (1, 2) , Hankui Hankz Zhuo (1, 2) , Andreas Herzig (3, 4) , Laurent Perrussel (3, 5, 6) , Peilin Chen (1, 2)

1 School of Data and Computer Science [Guangzhou]
2 SYSU - Sun Yat-sen University [Guangzhou]
3 IRIT-LILaC - Logique, Interaction, Langue et Calcul
4 CNRS - Centre National de la Recherche Scientifique
5 UT - Université de Toulouse
6 UT Capitole - Université Toulouse Capitole

Hierarchical Task Network (HTN) planning is showing its power in real-world planning. Although domain experts have partial hierarchical domain knowledge, it is time-consuming to specify all HTN methods, leaving them incomplete. On the other hand, traditional HTN learning approaches focus only on declarative goals, omitting the hierarchical domain knowledge. In this paper, we propose a novel learning framework to refine HTN methods via task insertion with completely preserving the original methods. As it is difficult to identify incomplete methods without designating declarative goals for compound tasks, we introduce the notion of prioritized preference to capture the incompleteness possibility of methods. Specifically, the framework first computes the preferred completion profile w.r.t. the prioritized preference to refine the incomplete methods. Then it finds the minimal set of refined methods via a method substitution operation. Experimental analysis demonstrates that our approach is effective, especially in solving new HTN planning instances.

Domaines: Intelligence artificielle [cs.AI]

Zhanhao Xiao, Hai Wan, Hankui Hankz Zhuo, Andreas Herzig, Laurent Perrussel, et al.. Refining HTN Methods via Task Insertion with Preferences. Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAi 2020), Feb 2020, New York, United States. pp.10009-10016, ⟨10.1609/aaai.v34i06.6557⟩. ⟨hal-02887591⟩