Sub goal oriented A* search

Abstract

Search is a well-studied paradigm of Artificial Intelligence (AI). The complexity of various search algorithms is measured in terms of space, and time to solve a problem. Blind search methods use too much space or too much time to solve a problem. Informed search algorithms such as Best First Search, A* overcomes these handicaps of blind search techniques by employing heuristics, but for hard problems heuristic search algorithms are also facing time and space problem. In the following study we present a new informed sub-goal oriented form of A* search algorithm. We call it “Sub-Goal Oriented A* Search (SGOA*)”, it uses less space and time to solve certain search problems which have well-known sub-goals. If we employ an admissible heuristic, then SGOA* is optimal. The algorithm has been applied to a group of fifteen puzzles.