It is important to compare the different mice on the same maze. To do this, the exhibitor's father changed the program to make a copy of each maze puzzle for each mouse. This means for each puzzle the mice solve the same maze.
For each puzzle the maze-solving engine will go through the maze by calling the Mouse's makeAMove method until it gets to the end of the maze. It also counts the number of moves that were made.
Randy always took fewer moves to solve the maze so it is better to have memory.
This project had a lot of data that had to be analyzed. Seven different mice were compared. Three different size mazes were used. Because the computer was used to gather the data, it was run overnight, resulting in over 5,000 different maze puzzles. So 105,000 data points were analyzed (5,000 mazes * 3 maze sizes * 7 different mice).
Here is a sample of the data that was run for the 50x50 maze. The complete set of data is 5,000 lines long (over 100 pages … just for the 50x50 maze). Each line shows the number of moves it took for the mouse to solve each maze puzzle.
A chart that plots these data points is not clear.
We definitely needed to find another way to look at the data.
Here are the averaged results for the 50x50 maze:
Here are the averaged results for the 16x24 maze:
Here are the averaged results for the 10x10 maze:
These summarized charts show a trend that repeats over and over again.
Here are the averages for the 50x50 maze:
Here are the averages for the 16x24 maze:
Here are the averages for the 10x10 maze: