Explicit context-based blackboards enhancing blackboard systems performance

Abstract

Inference engine in blackboard systems is better viewed as a process of matching potential knowledge sources with promising solution areas. To manifest this feature, a context-based blackboard is proposed as an explicit component to summarize, in terms of contexts, meta-level information about ordinary blackboards. It is so organized that a list of promising solution areas, a list of potential knowledge sources, and their best match can be proposed effectively and efficiently. The concept of context is properly elaborated and comprehensively used in the design of the blackboard. This approach improves the system performance by providing a balanced status feedback between potential knowledge sources and promising solution areas for system control and replanning. It also enhances the system performance by supporting viewpoint reasoning to allow reasoning on a single solution space from different aspects and various meta-level applications, e.g., viewpoint explanation, knowledge acquisition, etc.