Generic Multilayer Architecture of Knowledge-Based System for Structured knowledge representation.


Knowledge plays crucial role in routine to spectacular and complex problem solving activities in this knowledge era. Knowledge-based system utilizes knowledge gathered from multiple experts/sources for such problem solving and offer many advantages including documentation and long term storage of knowledge, in cost effective way. Inspite of the advantages, it is very difficult to develop, use, and maintain the knowledge-based system. One of the reasons for the difficulty is large volume of knowledge base having heterogeneous, unstructured, and continuously changing knowledge. Hence, structuring of knowledge base becomes trivial task to manage volume and ease of use. This paper introduces multilayer knowledge-based system architecture illustrated with an example where the knowledge base is structured into multiple partitions. The work presents generic architecture of multilayer knowledge-based system and partitions of knowledge base such as control and domain layers, control heuristics, and detailed design to access the layers. To demonstrate the working of the proposed architecture, a case of business advisory system for integrated rural development is presented with detail design, sample rules, structure of the content master repository, and interface repository. The paper concludes by discussing advantages of the proposed architecture and future enhancements in the field.

Journal of Hybrid Computing Systems, Vol. 3, Issue 1, 2010.