CONSTRUCTING EXPERT-SYSTEMS BY USING OBJECT-ORIENTED PROGRAMMING

The potential capabilities of intelligent systems led in recent years to the stormy development of means for constructing expert systems (ESs). The first such means, the so-called empty expert systems (EMYCIN, KES, ROSIE, etc.), reflected the ''power in knowledge'' principle explicitly formulated in [1]. Specialized problem-oriented knowledge plays a definitive role in modeling the expert behavior in this case. The first means of this type had, as a rule, a built-in inference mechanism which realizes general search methods in production systems. The requirement for flexibility and universality has become absolute in the next generation of means, i.e., the LOOPS, Knowledge Craft, KEE, and ART tool environments. The inference mechanisms are created directly in the development of the expert system on the basis of a collection of functional blocks. This makes it possible to take into account the specific nature of expert reasoning in various problem areas lo the greatest possible extent, albeit at the price of large resource and time expenditures.