Acquiring planning domain models using LOCM

The implemented LOCM algorithm, the assumptions that it is based on, and an evaluation using plans generated through goal-directed solutions, through random walk, and through logging human-generated plans for the game of freecell are described.