Learning Hierarchical Task Models from Input Traces. (22nd July 2014)

• Record Type: Journal Article
• Authors: Hogg, Chad; Muñoz‐Avila, Héctor; Kuter, Ugur
• Is Part Of: Computational intelligence. Volume 32:Number 1(2016)
• Page Start: 3
• Page End: 48
• Publication Date: 2014-07-22
• Subjects: HTN planning -- automated planning -- machine learning; Artificial intelligence -- Periodicals; Computational linguistics -- Periodicals; 006.3
• DOI: 10.1111/coin.12044
• Languages: English
• ISSNs: 0824-7935

Abstract

Abstract : We describeHTN‐MAKER, an algorithm for learning hierarchical planning knowledge in the form of task‐reduction methods for hierarchical task networks (HTNs).HTN‐MAKER takes as input a set of planning states from a classical planning domain and plans that are applicable to those states, as well as a set of semantically annotated tasks to be accomplished. The algorithm analyzes this semantic information to determine which portion of the input plans accomplishes a particular task and constructs task‐reduction methods based on those analyses. We present theoretical results showing thatHTN‐MAKER is sound and complete. Our experiments in five well‐known planning domains confirm the theoretical results and demonstrate convergence toward a set of HTN methods that can be used to solve any problem expressible as a classical planning problem in that domain, relative to a set of goal types for which tasks have been defined. In three of the five domains, HTN planning with the learned methods scales much better than a modern classical planner.