mnmnc/machinelearning_maze

Passing through maze with Tremaux algorithm and Dead End Filler.

MachineLearning_Maze

Passing Through Maze

Input

Input data is provided in two files:

• maze_128x128
• maze_16x16

Both contain coordinates for each wall in a maze.

Input format:

Two first are x and y coordinates. Last one is an instruction how to draw a wall.

0 0 0 - means draw in 0,0: one wall down and one wall to the right.    ┌
0 0 1 - means draw in 0,0: only wall to the right                      ─
0 0 2 - means draw in 0,0: only wall down                              │
0 0 3 - means draw in 0,0: no wall at all

Output for Tremaux:

Time spent solving: 0.11201095581054688 seconds
Iterations counted: 14641
Path lenght: 5547
Dead ends lenght: 4547

Output for Dead End Filler

Long processing time.

Time spent solving: 119.60996007919312 seconds
Iterations counted: 3700
Path length: 0
Dead ends length: 27218

Bonus round

Tremaux without randomness in path picking step (always picks first one - similar to Wall Follower, although its capable of leaving loops)