The Computational Complexity of Probabilistic Planning

TL;DR

This paper analyzes the computational complexity of probabilistic planning, revealing that various planning problems are complete for multiple complexity classes and introducing a new NP^PP-complete problem, E-MAJSAT.

Contribution

It characterizes the complexity of different probabilistic planning problems and introduces E-MAJSAT, a new problem that generalizes Boolean satisfiability for probabilistic computations.

Findings

Planning problems are complete for classes like PL, P, NP, co-NP, PP, NP^PP, co-NP^PP, PSPACE.

The complexity varies with plan type and definition of plan value.

E-MAJSAT is a new NP^PP-complete problem relevant for probabilistic reasoning.

Abstract

We examine the computational complexity of testing and finding small plans in probabilistic planning domains with both flat and propositional representations. The complexity of plan evaluation and existence varies with the plan type sought; we examine totally ordered plans, acyclic plans, and looping plans, and partially ordered plans under three natural definitions of plan value. We show that problems of interest are complete for a variety of complexity classes: PL, P, NP, co-NP, PP, NP^PP, co-NP^PP, and PSPACE. In the process of proving that certain planning problems are complete for NP^PP, we introduce a new basic NP^PP-complete problem, E-MAJSAT, which generalizes the standard Boolean satisfiability problem to computations involving probabilistic quantities; our results suggest that the development of good heuristics for E-MAJSAT could be important for the creation of efficient algorithms for a wide variety of problems.