Focusing Attention in Anytime DecisionTheoretic Planning
Peter Haddawy Decision Systems and Artificial Intelligence Lab Department of Electrical Engineering and Computer Science University of WisconsinMilwaukee Milwaukee, WI 53201 email@example.com
Any anytime algorithm used for decisionmaking should have the property that it considers the most important aspects of the decision problem first. In this way, the algorithm can first eliminate disas trous decisions and recognize particularly advan tageous decisions, considering finer details if time permits. We view planning as a decisionmaking process and discuss the design of anytime algo rithms for decisiontheoretic planning. In particu lar, we present an anytime decisiontheoretic plan ning algorithm that uses abstraction to focus atten tion first on those aspects of a planning problem that have the highest impact on expected utility. We discuss control schemes for refining this behav ior and methods for automatically creating good abstractions.