An Agent Architecture for Continuously Evolving Large-Scale Knowledge Bases Systems

In order for Knowledge-Based Systems to continue progressing into the mainstream of computing technology, they must become capable of addressing large and complex problems. Conventional systems are achieving this progress by moving to network infrastructure models, distributed systems, and finer granularity that is object-oriented. Analogous architectures are needed for Knowledge-Based Systems so boundaries can be crossed and new knowledge-based modules can be added as a system evolves. Our goal is to develop an intelligent agent architecture for continuously evolving large-scale knowledge-based, as well as hybrid, systems. The objectives are to show the feasibility of the architecture by designing the agent constructs that are at the core of the concept. This will be accomplished by building on our fault-tolerant agent work with enhancements to address the large-scale knowledge base requirements. Our proposed agent architecture is an enabling technological advance that can act as a catalyst for the evolution of large-scale knowledge-based systems and conventional and hybrid systems. The commercial potential comes from its applicability, especially if licensed, to the implementation of complex government systems (e.g., C3I), as well as complex systems for the medical, financial, and educational markets.