Maze Makers: Modified Cylog Maze Program for Tabbs Feature

[Boleeman]

"I honestly thought I was being a massive Graphics forum spammer and had a perception that I was being a bit of a pain, being too responsive to the "Post" button."

Thanks for your kind words TBMan and Phoenix.

[Boleeman]

Two posts before this one you can download the modified Max limits Laze source.

With the Laze Mazes program I experimented by changing the scaling slide Max limits and was pleasantly surprised that the maze creation was quite fast (almost instantaneous).

The solution path finding was quite slow (took about 12 minutes to solve). I was watching it work and sometimes it made some obvious wrong pathway choices (I could see the correct pathway but the program needlessly went down obvious dead ends and then backed up to the correct path).

I wondered if a different (more efficient) pathfinding method could be implemented for these larger mazes? The maze creation is quite fast but the solution finding is extremely slow.

I have a JavaScript Maze Generator (found on GitHub) that can make large mazes fast and it also solves them just as fast. Something like this is needed for the Lazarus pathfinder mechanism.

Lastly looking at the extreme sized Lazes maze, sometimes there is a vertical bias, sometimes horizontal and sometimes zigzag. So at very large sizes the Lazes program seems to have a mixture of all 3 biases. At the CSharp Helper Website ( http://www.csharphelper.com/howtos/howto_weighted_maze.html )  Rod Stephens made a maze with topographic features that uses a random spanning tree algorithm to make biased mazes with vertical, horizontal, age, zigzag and straight biased modes. I've known Rod for a long time and admire his CSharp skills (although my CSharp suck but I was getting a bit better at it). I had gotten a lot better at doing Lazarus Graphics programming). Rod has certainly made a lot of interesting programs.

[Boleeman]

Cylog's LazMaze V3

Found it on SourceForge and compiled it to 64Bits (was 32Bit).

Can left click for start and right click for end positions when doing the pathfinding.

Not sure how to toggle (hide/show) the solution without creating a new maze?

[Thaddy]

Maze solutions - in a true maze - are easy. Always go left or always go right. Do not mix.
For the shortest way, there is technically no direct answer without knowing the solution beforehand. (That would be Dijkstra's).
I already gave you that answer?

And yes, while the left/right paradigm always hold, maze solutions are otherwise computationally expensive (and fun).
But solving a maze isn't: always left or always right. Without prior knowledge you can't compute Dijkstra's shortest path.

[speter]

You may like to have a look at this thread; it includes a recursive (maze) pathfinder.

cheers
S.

[Boleeman]

Thanks.

Actually I thought of just copying the maze to a temporary bmp (where it can remain untouched until New Maze button is clicked) and then different solutions can be found by clicking in different regions of the original canvas of the maze.

speter, I will look at your standard recursive pathing algorithm code as well.

Thanks Thaddy and speter.

[Boleeman]

And the updated Cylog Lazmaze maker.

Tabbed to play around with solution (in order to get more difficult path finds).

Left and right mouse click to set the starting point and the end point of the path finder.

Also added a Save to Png feature that saves both mazes with given sub-names.


Enjoy.

[Boleeman]

Recently came across an article about using CSharp to make "Mazed Progress Bars".

https://www.codeproject.com/articles/The-Amazing-ProgressBar-Control#comments-section

The progress bars kinda LOOK COOL.

Also working on a Cretan Labytinth (uses 1 line only and threads it).  Not a maze but makes amazingly tight paths. Nice!