A hierarchical blackboard architecture for distributed AI systems

A framework for distributed problem solving should support the incremental development and integration of problem solvers into the system in agreement with the principles of stepwise refinement and functional decomposition. The organizational structure of the framework should guide the mapping of each of the functional areas onto agents and its communication structure should reflect this decomposition. The authors describe a hierarchical multi-blackboard architecture, called the Hierarchical Blackboard System (HBBS) which satisfies these requirements. It serves as an integration architecture for heterogeneous expert systems. Important features of the system, such as the hierarchy concept, communication strategy and configuration management of the system at run-time are discussed. A working prototype HBBS for distributed CAD/CAE of communication systems is described