Approximate reasoning and fuzzy evaluation in code compliance checking

Abstract

This paper deals with the problem of developing intelligent tools which would support the design system by automated checking various design criteria. These criteria contain both building codes and customer requirements. The paper extends the logic-based reasoning methods used previously in computer-aided visual design to a fuzzy classification. The fuzzy classification of the drawings representing early design solutions is based on approximate reasoning, where different types of criteria are considered. Due to this classification the system is able to evaluate the correspondence of potential solutions to the project specification and check their code compliance. The approach is illustrated by example of designing layouts of a small office.