CaseNet: A Distributed Case-based Reasoning Network

. In a network of CBR systems, each system may have different capabilities and availability for solving any particular problems. It is expected that by integrating multiple CBR systems into a distributed environment, it is possible to can provide better solutions for problems more efficiently. CaseNet is a distributed CBR network which integrates multiple CBR systems as well as human experts, and employs a broker to select appropriate CBR systems and human experts for better solution of problems according to their changing capabilities and availability. This paper describes its architecture and related learning and scheduling algorithm. We show that CaseNet is a flexible distributed network on which human and CBR systems can be easily added or deleted at any time. A novel feature of CaseNet is its ability to incorporate solver capabilities and availability that are dynamically changing in its scheduling algorithm, so that the overall system performance is kept at near optimal. We highl..