Abstract:We study the complexity-theoretic boundaries of tractability for three classical problems in the context of Hierarchical Task Network Planning: the validation of a provided plan, whether an executable plan exists, and whether a given state can be reached by some plan. We show that all three problems can be solved in polynomial time on primitive task networks of constant partial order width (and a generalization thereof), whereas for the latter two problems this holds only under a provably necessary restriction to the state space. Next, we obtain an algorithmic meta-theorem along with corresponding lower bounds to identify tight conditions under which general polynomial-time solvability results can be lifted from primitive to general task networks. Finally, we enrich our investigation by analyzing the parameterized complexity of the three considered problems, and show that (1) fixed-parameter tractability for all three problems can be achieved by replacing the partial order width with the vertex cover number of the network as the parameter, and (2) other classical graph-theoretic parameters of the network (including treewidth, treedepth, and the aforementioned partial order width) do not yield fixed-parameter tractability for any of the three problems.
What problem does this paper attempt to address?
This paper mainly discusses the computational complexity bounds of three classic problems in Hierarchical Task Network (HTN) planning: plan validation, existence of executable plans, and state reachability. The focus of the research is to determine under which special cases these problems can be effectively solved and when the effective solutions for simple cases can be generalized to general cases in more general task networks.
Firstly, the paper points out that these three problems can all be solved in polynomial time on primitive task networks with constant-width partial orders. However, the effectiveness of the solution is limited to cases where the state space is subject to specific constraints for the last two problems. Next, the authors obtain an algorithmic metatheorem and corresponding lower bounds to identify the conditions under which the polynomial-time solvability results from primitive task networks can be extended to general task networks. Furthermore, they analyze the parameterized complexity of these three problems, showing that by replacing partial order width with vertex cover number as the parameter, fixed-parameter tractability can be achieved for all three problems, while other classic graph theory parameters such as tree width and tree depth cannot bring fixed-parameter tractability to these three problems.
The paper further demonstrates that on networks with only chains or anti-chains, these three problems are all solvable in polynomial time, implying the existence of solvability "pockets" in the intractability of computational complexity. Then, they prove that these special cases actually represent a spectrum of solvability special cases, namely, on networks with finite "generalized partial order width" when isolated elements are ignored, these three problems are solvable in polynomial time.
For primitive task networks, the paper also presents an algorithmic metatheorem describing how to extend the effectiveness results for the primitive task networks to the general case in compound task networks when the number of compound tasks, the maximum network size of task decomposition, and the maximum depth of decomposition are all limited by constants. At the same time, the paper points out that solvability cannot be guaranteed if any of these restrictions is relaxed.
Lastly, the paper studies these three problems from the perspective of parameterized complexity and proves that generalized partial order width is not applicable to fixed-parameter tractability, but fixed-parameter tractability can be achieved by setting the parameter as the vertex cover number of the network. For compound HTN, they extend the algorithmic metatheorem regarding polynomial-time solvability to the scenario of fixed-parameter tractability.
Overall, this paper fills the gap in the theoretical understanding of computational complexity in HTN planning problems and reveals the solvability and complexity boundaries under different structural parameters.