Different Maze Types Implementation

[romandykyi]

Hi, I'm the author of the Labyrinthian library, which specializes in step-by-step generation of various graph-based maze types (e.g., delta, theta, upsilon) using different algorithms. Recently, I integrated your Origin Shift algorithm into my library, and users can now visualize the generation process in this online demo.

In my implementation, I start by selecting a cell (random by default) and perform a BFS from it to create the initial spanning tree (perfect maze). Then, I apply n iterations (n = 10 * mazeCellsCount by default, as per your suggestion in your video).

Your algorithm has truly impressed me. Unlike other graph-based algorithms like Kruskal or Wilson that generate a random spanning tree from scratch, Origin Shift operates on an existing spanning tree and dynamically modifies it. I also find it particularly innovative that directed graphs can be utilized effectively for maze generation, I haven't encountered a similar approach before.

However, Origin Shift does have its limitations. Occasionally, it fails to fully traverse the graph, especially with Delta mazes, even with 10 * mazeCellsCount iterations:

Despite this, I believe your algorithm has significant potential for enhancement or as a valuable secondary post-processing tool.

[CaptainLuma]

Hi Roman, thank you so much for taking interest in Origin Shift. The Labyrinthian library is really great! I love the idea of a site which catalogs all maze generating techniques and shows visual demo's for each, and I think the Labyrinthian is a wonderful execution of that idea. It is so cool to have my algorithm featured on there. The problem of the algorithm not always traversing the whole grid is a problem that a few people have pointed out. I believe it has to do with my suggested iteration equation of "i = 10a" being inaccurate. I basically eyeballed the equation and only tested it out with smaller mazes. A more accurate equation could be found mathematically, though I am not really skilled enough in math to do this. But if I do discover a more accurate equation, then I will let you know. Thank you again for adding Origin Shift to your site. It is an honor! - Luma

…

On Tue, 2 Jul 2024 at 16:35, Roman Dykyi ***@***.***> wrote: Hi, I'm the author of the Labyrinthian <https://github.com/romandykyi/Labyrinthian> library, which specializes in step-by-step generation of various graph-based maze types (e.g., delta, theta, upsilon) using different algorithms. Recently, I integrated your Origin Shift algorithm into my library, and users can now visualize the generation process in this online demo <https://romandykyi.github.io/LabyrinthianDemo/>. In my implementation, I start by selecting a cell (random by default) and perform a BFS from it to create the initial spanning tree (perfect maze). Then, I apply n iterations (n = 10 * mazeCellsCount by default, as per your suggestion in your video). Your algorithm has truly impressed me. Unlike other graph-based algorithms like Kruskal or Wilson that generate a random spanning tree from scratch, Origin Shift operates on an existing spanning tree and dynamically modifies it. I also find it particularly innovative that directed graphs can be utilized effectively for maze generation, I haven't encountered a similar approach before. However, Origin Shift does have its limitations. Occasionally, it fails to fully traverse the graph, especially with Delta mazes, even with 10 * mazeCellsCount iterations: Triangular.maze.size.50.png (view on web) <https://github.com/CaptainLuma/New-Maze-Generating-Algorithm/assets/94003504/01f862e3-866a-4b4f-8d48-5ee9455d3c95> Despite this, I believe your algorithm has significant potential for enhancement or as a valuable secondary post-processing tool. — Reply to this email directly, view it on GitHub <#5>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BFK4FWSMDVUVAUPRKMC4GULZKMTLDAVCNFSM6AAAAABKIM6BRKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGM4DOMRXGA3DIMI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>