Figure 7: Maze environment information acquirement.
The subgraphs in the top row represent the situations of maze exploration by the rat. The parts that are covered by gray denote the units which the rat has not visited yet, while the uncovered parts represent the units the rat has travelled. In the bottom row, the subgraphs show the environment information of the maze that the computer has got. As the number of the discovered decision points increasing in the top row, the relative state was added into the existing state set (i.e. the state s1 was added into the state set when the first decision point was found) initiated as {the start state s0, the final/target state sf}. And when the rat passed the first decision point, the action set was configured as {correct, wrong}.
