An open intelligent information systems architecture
Rajeev Kaula · Medical Entomology and Zoology · 1999
The conventional wisdom in intelligent information processing system development is that systems should be developed with global consistency and completeness. This viewpoint has increased the complexity of systems development. Hewitt and deJong (1984) suggested an alternative approach for intelligent systems development called the open-systems approach. The open-systems approach envisages intelligent systems consisting of many small autonomous intelligent subsystems which cooperate by communicating with each other during problem solving. In such an environment there is no global schema, consistency or completeness. However existing information processing conceptual architectures like composite database, federated database, production systems and blackboard systems architectures have limitations in their use toward realizing open intelligent information systems. The dissertation proposes a new conceptual architecture called an open intelligent information systems (OIIS) architecture for the design of open intelligent information systems. The architecture has many components. First there is a user interface through which the user interacts with the system. Secondly there is the provision for multiple autonomous intelligent subsystems. Thirdly there is a concept dictionary, which provides an index of specific knowledge handled by various autonomous intelligent subsystems. The mode of communication within the OIIS is based on the speech act theory of language proposed by Searle (1969). However speech acts have been modified and called message acts, and implemented in the communications dictionary component. Autonomous intelligent subsystems directly communicate with each other using message acts (communications dictionary) after making inquiries to the concept dictionary during problem solving. A business application prototype is developed in the Prolog language to test the architecture's working. The development of the OIIS architecture in the dissertation has major implications for the future development of intelligent systems, apart from providing new directions for further research.