The Heuristic of George Polya and Its Relation to Artificial Intelligence
CARNEGIE-MELLON UNIV PITTSBURGH PA DEPT OF COMPUTER SCIENCE
Pagination or Media Count:
Polyas fundamental work in heuristic is well known and well regarded in artificial intelligence. However, no one has built seriously on his work, e. g., by constructing programs that make use of his heuristic. This paper attempts to understand why this might be the case. First, an attempt is made to characterize the nature of Polyas heuristic. Then six theses are put forward that might account for the failure of his work to have a major impact. Three are easily discarded, but three are serious candidates that the essential heuristic knowledge is not captured in Polyas work that the emphasis on learning in Polyas heuristic is beyond the current art in artificial intelligence and that the use of auxiliary problems is beyond the current art. This last thesis is explored in detail in the remainder of the paper. Some interesting concepts emerge, particularly the notion of object-centered problem space and the contrast between tame and wild subproblems.