We show several natural questions about hinged dissections of polygons to be PSPACE-hard. The most basic of these is: Given a hinged set of pieces and two configurations for them, can we swing the pieces on the hinges to transform one configuration to the other? We also consider variants in which the configurations must be convex, the placement of hinges is not specified, or the configurations are not supplied, but just the target shapes. We show all of these variants to be PSPACEhard, via a reduction from Nondeterministic Constraint Logic [4].
Robert A. Hearn, Erik D. Demaine, Greg N. Frederic