Hierarchical Task Network Planning with Task Insertion and State Constraints

Abstract

We extend hierarchical task network planning with task insertion (TIHTN) by introducing state constraints, called TIHTNS. We show that just as for TIHTN planning, all solutions of the TIHTNS planning problem can be obtained by acyclic decomposition and task insertion, entailing that its planexistence problem is decidable without any restriction on decomposition methods. We also prove that the extension by state constraints does not increase the complexity of the plan-existence problem, which stays 2-NEXPTIME-complete, based on an acyclic progression operator. In addition, we show that TIHTNS planning covers not only the original TIHTN planning but also hierarchyrelaxed hierarchical goal network planning.

Correcting Hierarchical Plans by Action Deletion

“…Let us now formalize the constraints C used in the set of available decomposition methods M . They are the ones available in the original HTN formalization (Erol, Hendler, and Nau 1996), but have been re-introduced in other hierarchical planning formalizations as well (Xiao et al 2017). The constraints can be of the following three types, where the first is also known as an ordering constraint and the latter two are essentially state constraints:…”

Efficient Optimal Approximation of Discrete Random Variables for Estimation of Probabilities of Missing Deadlines

“…In light of this, we tested the accuracy of the repetitive-linApprox to see how it performs against the tools of (Cohen, Shimony, and Weiss 2015;Cohen, Grinshpoun, and Weiss 2018) using their benchmarks. These benchmarks are taken from the area of task trees with deadlines, a sub area of the well-established hierarchical planning (Dean, Firby, and Miller 1988;Alford et al 2016;Xiao et al 2017).…”

HTN-like solutions for classical planning problems: An application to BDI agent systems

“…Unlike their work, ours provides an account of abstraction that also takes into account the notion of redundancy, which we consider relevant when classical planning is in the context of HTN-like available knowledge. The strand of work on HTN planning with "task insertion" (TIHTN) [24,48,3,2] is related to [27] with the main difference being that the planning problem in the latter (and in our work) contains a goal condition, whereas the planning problem in TIHTN planning contains an initial task network.…”