Overview
Dr Ava Clifton is a computer scientist and researcher specialising in automated planning, Boolean Satisfiability (SAT), and AI search paradigms.
She completed her undergraduate studies (2016–2019) and earned her PhD (2020–2025) from the College of Systems and Society at the Australian National University (ANU). Her doctoral thesis, titled “Advancing Property Directed Reachability for Classical and Fully Observable Nondeterministic Planning”, focused on expanding Property Directed Reachability (PDR)—a powerful search methodology originating in hardware and software model checking—into classical and non-deterministic planning. She has also invented PDR algorithms for the Fully Observable Non-Deterministic planning problem.
Her research contributions include developing multi-step SAT algorithms and parallel PDR frameworks (such as Parallel State PDR) to accelerate planning under uncertainty. She has published her findings at leading automated planning and artificial intelligence venues like ICAPS and KR, and currently serves as a Postdoctoral Researcher in the School of Computing at ANU.
parallel-pdr
Links: Code (GitHub) KR Parallel PDR and Decompositional PDR Paper (2023)
I invented the parallel-pdr planning system built using the Dagster skeleton. It implements several parallel computing strategies, featuring a novel distributed PDR algorithm designed to solve problems described in PDDL.
FOND-pdr
Links: Code (GitHub) AJCAI (2026)
I extended PDR for the first time to tackle the Fully Observable Non-Deterministic (FOND) planning problem, which generalises classical planning by allowing actions to have multiple possible outcomes chosen nondeterministically at execution.