Overview over Rush Hour Research
Rush Hour is a two-dimensional puzzle game invented by Nob Yoshigahara in the late 1970s and which is now sold by ThinkFun.
The story of the game is that you have parked your car (a red car) in a parking lot and now want to leave. On the parking lot are other cars which block the exit and you have to sequentially move the cars, s.t. you can reach the exit with your car.
Here you can see an image of the game board:
The goal is to reach the exit at the right border of the board with your red car. Cars can only be moved longitudinally. Thus, in this simple example you first have to move the red car (1) forwards, then move the blue car (4) upwards, move the yellow car (3) to the left so that you can move the green car (2) downwards. Then the way to the exit is free and we can move the red car out.
Of course, this is just a simple example and there are much much harder problems. The hardest Rush Hour problems you can generate from the standard game set are problems having an optimal solution of over 90 moves.
To study the human behavior in this planning domain we conduct behavioral experiments in which participants have to solve Rush Hour problems in a computer version of the game. The program records the participants' actions and the response times which we then use to get a detailed overview of his or her behavior.
We integrated the recorded actions in game graphs to visualize the state transitions. This allows us to have a general overview of frequent transition paths and to identify difficult situations. Further we can extract sub paths from single participants to scale the analysis down to individual behavior.
Because the Rush Hour game has a large state space and participants do not perform the same actions, the transition graphs become very large.