Equivalence between the paper’s supported temporal-planning subset and the verified validator’s semantics

Prove the equivalence between the temporal-planning semantics used by the paper’s timed-automata encoding for the supported subset of PDDL features and the semantics implemented by the formally verified temporal plan validator of Abdulaziz and Koller (AAAI 2022), i.e., show that for every ground temporal planning problem within the overlapping feature set, plan validity coincides under both semantics.

Background

The paper develops a formally verified reduction of temporal planning problems to networks of timed automata and aligns its semantics to facilitate reasoning and certification. While the reduction currently supports a subset of PDDL features, the authors relate this subset to the features handled by a previously verified validator for temporal planning (Abdulaziz and Koller, AAAI 2022).

They explicitly conjecture that their supported subset corresponds equivalently to the validator’s semantics and state an intention to formally verify this equivalence. Establishing this result would strengthen end-to-end trust in their certification pipeline by ensuring semantic compatibility across tools.