A new algorithm for heuristic search which uses hints

Abstract In computer aided decision making an algorithm to determine the next step (i.e. the decision) must be formulated. However, often we are not able to express our knowledge of the problem algorithmically and the decision making is based on heuristic search, which is usually guided by some cost evaluation function — heuristic function. But quite often even this heuristic function cannot be formulated ideally (due to the lack of our knowledge of the problem). When such a function is used by the “classical” A ∗ algorithm [Nilsson, 1971], whole subspaces not containing the solution path can be searched in vain. We propose a new search guiding algorithm which allows one to specify also other forms of heuristic knowledge to reduce the search. We analyze the new algorithm, state and prove its basic properties and compare it with known algorithms.