This paper describes the topological aspect of a logic-based, artificial intelligence approach to formalising the qualitative description of spatial properties and relations, and reasoning about those properties and relations. This approach, known as RCC theory, has been under development for several years at the University of Leeds. The main rationale for this project is that qualitative descriptions of spatial properties and relationships, and qualitative spatial reasoning, are of fundamental importance in human thinking about the world: even where quantitative spatial data (...) are most important, they must be attached to the components of a perceived spatial structure if we are to make use of them. RCC theory covers other qualitative aspects of spatial description and reasoning, but the topological properties and relations of spatially extended entities are fundamental to our work. The topological formalisms used by mathematicians are, in general, not well suited to the task of formalising the kinds of ‘common-sense’ or ‘everyday’ qualitative spatial description and reasoning which are our primary interest. Nevertheless, we must come to grips with the concepts of topology as practised by mathematicians if we are not to risk constantly ‘reinventing wheels’. (shrink)
Based on the proceedings of a 1985 conference held in the U.K., this volume embraces most of the important concerns in AI today, emphasizing common techniques and methodologies rather than applications. Topics covered include building efficient computational logic, planning and design, the representation of uncertain knowledge, user modelling, and psychological and philosophical issues. Papers on perception, theorem proving, expert systems, robotics, and data bases are also included. Each section is preceded by an introduction which draws comparisons between various papers.