Case-Based Reasoning for Deployment Transport Planning
Final rept. Feb 1991-Feb 1994
CHICAGO UNIV IL ARTIFICIAL INTELLIGENCE LAB
Pagination or Media Count:
In the report we describe work done in the application of case-based reasoning CBR techniques to large scale planning problems. We report on systems that demonstrate the role of CBR in four areas of planning CBR as Gatekeeper. The core idea of cased-based reasoning is the re-use of plans by relating past experience to present problems. This makes CBR ideal as the gatekeeper between users and specialized problem solvers in two ways the retrieval of past solutions, and the retrieval of past solution strategies. CBR as Guide. With the complex constellation of problem solvers and information sources available in large-scale planning, guiding users among available resources becomes of crucial importance. By retrieving solution strategies, case-based reasoning can provide structure even in the absence of past solutions. CBR as editor. Specialized problem solvers make use of specialized problem representations, incomprehensible to those unfamiliar with the technology. We demonstrate that case-based reasoning can serve as an editor of this specialized jargon, presenting results and options in terms that are clear and useful. CBR as Opportunist. It is an axiom of large-scale planning that solutions must always be partial and approximate, to be further developed over time.