Jendrik Seipp, WASP Fellow and senior associate professor at Linköping University received the ICAPS 2025 Influential Paper Award, awarded annually to a planning paper at least ten years old that has had a significant impact on the field of automated planning and scheduling. His 2013 paper “Counterexample-guided Cartesian Abstraction Refinement”, co-authored by Malte Helmert, introduced a novel abstraction refinement technique that has since become a standard tool in the field, inspiring numerous follow-up works and advancements in planning research.
The method applies counterexample-guided abstraction refinement (CEGAR) to planning: it starts with an extremely coarse, simplified model of the planning problem and finds an optimal “plan” in that abstract model. If that abstract plan fails in the real problem, the failure (counterexample) tells the algorithm how to refine the model so the same mistake can’t happen again. A key innovation was a novel class of Cartesian abstractions that allow very fine-grained refinement. In practice, Seipp’s implementation could perform tens of thousands of such refinements in a few minutes. The resulting heuristics provided much stronger cost estimates than previous abstraction techniques of the same size.
The award was presented at the International Conference on Automated Planning and Scheduling (ICAPS) 2025, held in Melbourne, Australia, from November 9-14, 2025. ICAPS is the premier forum for automated planning research (CORE Rank A*).
Published: December 5th, 2025
[addtoany]