How Humans Search Varying-Knowledge Environments: Solving Imperfect Information Mazes

Abstract

In our experiment, we studied the searching sequence of humans, i.e. how humans solve the game tree. We created a controlled environment of maze game to simulate the situation where the player is not allowed to observe the entire map freely. We use "fog-block" to cover the intersection of a maze. Thus, the maze becomes an "imperfect information maze." We give additional information to hint on the path that the player should choose. Then we observed the change in players' behavior. We found that hints affect searches making it best fit path towards hint.