Heuristics as Invariants and its Application to Learning
We present a characterization of heuristic static evaluation functions which unifies their treatment in single-agent problems and two-person games. The central thesis is that a useful heuristic function is one which is invariant along a solution path. This local characterization of heuristics can be used to predict the effectiveness of given heuristics and to automatically learn useful heuristic functions for problems.
Unable to display preview. Download preview PDF.