A software architecture for knowledge-based systems

Abstract

The paper introduces a software architecture for the speci®cation and veri®cation of knowledgebased systems combining conceptual and formal techniques. Our focus is component-based speci®cation enabling their reuse. We identify four elements of the speci®cation of a knowledgebased system: a task de®nition, a problem-solving method, a domain model, and an adaptor. We present algebraic speci®cations and a variant of dynamic logic as formal means to specify and verify these dierent elements. As a consequence of our architecture we can decompose the overall speci®cation and veri®cation task of the knowledge-based systems into subtasks. We identify dierent subcomponents for speci®cation and dierent proof obligations for veri®cation. The use of the architecture in speci®cation and veri®cation improves understandability and reduces the eort for both activities. In addition, its decomposition and modularisation enables reuse of components and proofs. Therefore, a knowledge-based system can be built by combining and adapting dierent reusable components.

A conceptual architecture for semantic web enabled web services

“…Such adapters enable reusable descriptions of objects and make it possible to combine objects that differ in their syntactical input and output descriptions. [4] introduced the concept of an adapter in architectural descriptions of knowledge-based systems to (1) de-couple different elements of this model, (2) encapsulate these different elements and (3) explicitly model their interactions. Work on software architectures describes systems in terms of components and connectors that establish the proper relationships between the former (cf.…”

A Formal Approach to Designing Reliable Advisory Systems

“…The basic purpose of an advisory system is to assist the end-user in identifying suitable solutions to complex, unstructured problems [1][2][3][4][5][6][7][8][9][10]. In decision-making, an un-structured problem is one that is characterised with contextual uncertainty, where there are no definite processes in place for predictably responding to a problem that is, well-defined actions that do not necessarily lead to predictable outcomes [2].…”

“…The basic purpose of an advisory system is to assist the end-user in identifying suitable solutions to complex, unstructured problems [1][2][3][4][5][6][7][8][9][10]. In decision-making, an un-structured problem is one that is characterised with contextual uncertainty, where there are no definite processes in place for predictably responding to a problem that is, well-defined actions that do not necessarily lead to predictable outcomes [2]. As such, problems of this nature require an analysis of all available information in order to properly describe the problem and to attribute suitable and realistic actions that minimises risk and maximises the possibility of yielding a positive outcome [1,9].…”

“…From these applications, it is clear that designs of advisory systems are based on the illustrated architecture and perform according to one of two main styles. These are: (1) monitoring and evaluation and (2) diagnosis and recovery [2][3][4][5][6][7][8][9]. In the monitoring and evaluation style, advisory systems simply monitor data streams to identify statistical anomalies that may represent a potential problem or to identify predictive behaviour patterns.…”

An approach for knowledge interaction to improve the availability of KBS

“…On the contrary, Fensel first investigated the knowledge interaction in KBS, and proposed a kind of classical software architecture of KBS based on Adapter, called UPML (The Unified Problem-solving Method development Language) [8], [9]. This work mainly focused on the interaction relationships among knowledge components in reusable architecture level.…”