Landmark Generation in HTN Planning

Landmarks (LMs) are state features that need to be made true or tasks contained in every solution of a planning problem. They valuable source information and can exploited various ways. LMs have been used both classical hierarchical planning, but while there is much work the techniques less evolved. We introduce novel LM generation method for Hierarchical Task Network (HTN) show it sound incomplete. complete approach as hard co-class underlying HTN problem, i.e. coNP-hard our setting (while P). On widely benchmark set, finds more than twice number landmarks from literature. Though focus on generation, we newly discovered bear beneficial solvers.