next up previous
Next: Acknowledgements Up:

Deep Interoperation between Previous: Related Work

Summary, Discussion and Future Work


In order for distributed diagnostic expert systems to exchange information at a conceptually acceptable level of detail for a better individual performance, IPST have been proposed. Furthermore, a cooperation method based on specification sharing has been presented, which focuses on the difference in context of the correspondence between inference primitives of an originator and those of a recipient. The wrapper with conversion facilities has also been provided, using a common domain ontology. After putting them into an integrated deep interoperation environment between distributed diagnostic expert systems, the experiment results have shown that the interoperation works on the following systems similar to model-based diagnostic expert systems in real task-domains: a trouble-shooting expert system and an enterprise diagnosis expert system.

However, there are still several issues in expert systems deep interoperation at a semantic level. The two expert systems have many similar points. In the case of many heterogeneous distributed expert systems, for example, among design expert systems and diagnosis ones, the interoperation environment works well ? When the number of expert systems increases, the interoperation environment still works ? Perhaps, we must change a cooperation method, from direct communication to assisted coordination. Although a common domain ontology has been manually constructed in order to ease correspondence between them, how do we automatically make large scale ontologies available for the interoperation environment ?

We have already developed other types of expert systems, such as a scheduling expert system and a software process design support system. Therefore based on the work here, we are planning to approach the above-mentioned issues.

Daiki Kishimoto s0011
Sat Sep 28 20:16:55 JST 1996