Cashing in on Caching: An Architecture for Time-BoundedKnowledge-Based Problem Solving

摘要: Knowledge-based computing, in general, suffers from an inherent open-endedness that precludes its application time-bounded domains where answer must be computed within a stipulated time limit. We examine two-way improvement of the shortcomings: knowledge representation scheme provides easy access to relevant and thereby reduces search time, reasoning is algorithmic nature thus makes computational requirements meaningfully estimable. In this work, we offer cache-based architecture capable both storing different formats (e.g. rules, cases), invoking appropriate fit available computing time. The cache helps retrieving most pieces (not only exact matches) required for solving given problem. This relies on tactic, interpolation, can generate solution two near-matches way, solutions. illustrate design such resource allocation problems domain shortwave radio transmission evaluate performance observing imposed temporal bounds.