Earl D. Sacerdoti

AUTOMATIC PROBLEM SOLVING' For intelligent computers to be able to interact with the real world, they must be able to aggregate individual actions into sequences to achieve desired goals. During the last decade, a number of techniques have been developed for improving the efficiency of these strategies. The bulk of this paper consists of a description of the problem-solving strategies and a catalogue of tactics for improving their efficiency.This is followed by an attempt to'This is a slight revision of a paper presented at the Sixth International Joint Conference on Artificial Intelligence, Tokyo, Japan, August 20-24, 1979 The original vetsion was prepared while the author was with SRI International, Menlo park, California, suppotted by the Defense Advanced Research Projects Agency under constract N00039-79-C-0118 with the Naval Electronic Systems Command The general function of an automatic problem solving system, then, is to construct a sequence of actions that transforms one world model into another. There are three basic capabilities that a problem solving system must have. These are: 1. Management of State Description Models The information will not all be explicitly encoded, however, so a deductive engine of some sort must be provided to allow needed information to be extracted from a model.