Unplannability IPC Track
Christian Muise, Nir Lipovetzky · 2015
The majority of research in the field of automated planning focuses on the synthesis of plans for problems that are solv-able. We propose an IPC track to focus on the important and understudied area of unplannibility: proving that a plan-ning problem is unsolvable. We will focus on classical plan-ning problems, as methods for determining whether or not unplannability can have wider applications for classical plan-ning problems (e.g., recognizing and avoiding deadends in the state space) as well as solving planning problems with uncertainty (e.g., identifying when a deterministic approx-imation of the problem is unsolvable). The unplannability track follows similar contests in other fields; for example, the UNSAT track for the field of Boolean Satisfiability. In a sim-ilar vein, we hope that the introduction of an unplannability track will foster new innovation for techniques dedicated to identifying planning problems that cannot be solved.