Greedy algorithm in block world

Block world is a convenient representation for planning problems. There are many different algorithms of state space searching which use STRIPS representation for planning. Some of them search for solution through a space of world-states. In general to generate an optimal plan for block world instances, where the goal is a complete description of a set of stacks, is NP-hard. In the paper a greedy strategy used to serialize instance subgoals is presented. It is shown when such strategy can reduce generate optimal plan problem to polynomial time problem and still generates optimal solution. An example explain greedy algorithm in block world. The problem of greedy algorithms optimality is also discussed.