Abstraction search dates back to the early days of artificial intelligence. The central idea is to identify important aspects of a given problem, construct a solution outline that ignores less significant aspects, and use it to guide the search for a complete solution. The applications of this approach take different shapes, and their utility varies widely across systems and domains. The related analytical models give disparate efficiency predictions for specific abstraction techniques; however, researchers agree that a proper abstraction can enhance almost all search systems.
KeywordsPrimary Effect Inference Rule Component Graph Abstract Solution Logistics Domain
Unable to display preview. Download preview PDF.