Abstract
The purpose of this article is to present and analyse issues connected with the expert knowledge collection methodology in the decision support systems (DSS). The considerations concentrate on a conception of building an information system, based on an application of case-based reasoning method and reasoning based on approximate knowledge. The expert's knowledge is systematically collected in a case base. A mechanism of classical CBR and a logic model of the case base were described. It was assumed that the cases compared with regard to similarity are elements of tolerance space what considerably accelerates the retrieval of satisfying solutions. A local and global measure of case similarity is developed. The method can be used in complex tasks of image identification.