Reconstructing hard problems in a human-readable and machine-processable way

Abstract

This paper shows how a controlled natural language can help to reconstruct a logic puzzle in a well-defined subset of natural language and discusses how this puzzle can then be processed and solved using a state of the art model generator. Our approach relies on a collaboration between humans and machines and bridges the gap between a (seemingly informal) problem description and an executable formal specification. © 2008 Springer Berlin Heidelberg.