We introduce lower-bound certificates for classical planning tasks, which can be used to prove the unsolvability of a task or the optimality of a plan in a way that can be verified by an independent third party. We describe a general framework for generating lower-bound certificates based on pseudo-Boolean constraints, which is agnostic to the planning algorithm used. As a case study, we show how to modify the A* algorithm to produce proofs of optimality with modest overhead, using pattern database heuristics and hmax as concrete examples. The same proof logging approach works for any heuristic whose inferences can be efficiently expressed as reasoning over pseudo-Boolean constraints.

Abstract We present a generic tree-interpolation algorithm in the SMT context with quantifiers. The algorithm takes a proof of unsatisfiability using resolution and quantifier instantiation and computes interpolants (which may contain quantifiers). Arbitrary SMT theories are supported, as long as each theory itself supports tree interpolation for its lemmas. In particular, we show this for the theory combination of equality with uninterpreted functions and linear arithmetic. The interpolants can be tweaked by virtually assigning each literal in the proof to interpolation partitions (colouring the literals) in arbitrary ways. The algorithm is implemented in SMTInterpol.

Abstract The verification approach of Ultimate Automizer utilizes SMT formulas. This paper presents techniques to keep the size of the formulas small. We focus especially on a normal form, called CommuHash normal form that was easy to implement and had a significant impact on the runtime of our tool.