A distributed real-time knowledge-based system and its implementation using object-oriented techniques

Abstract

We present a distributed knowledge-based system, called HOPES, for real-time applications. System details of HOPES are described, including its knowledge sources, the multi-blackboard architecture, and communication strategies. It shows that real-time AI systems can benefit very much from using distributed and cooperative processing techniques. Implementation issues of the HOPES system are also addressed. Object-oriented techniques are used for the system development. Important features of object-orientation such as message passing, encapsulation, and inheritance are discussed, and it is shown that object-oriented methods can provide a powerful means for implementing a distributed knowledge-based system effectively and efficiently.<>