Random L Blocks on a Grid

Replies to This Discussion

Permalink Reply by David Rutten on January 5, 2017 at 2:40pm

This is surprisingly difficult. And my solution is pretty inefficient.

My approach was to first only deal with the central curves, not the outlines. Then generate all possible L-Shapes that reside on integer coordinates within some rectangular region. This list will be ordered, so we need to randomize it using a Jitter component.

Once we have a list of all possible L-shapes, we start testing each one to see if it intersects with any of the previous ones.* If it does, we do not add it to the final collection.

Once we have a list of True/False values (one for each possible shape) where True means it intersects and False means it doesn't, we can then filter the original list of orientations and transform our L-shaped outline onto each one.

* A problem here is that each shape is intersected with all previous shapes, even the ones that were rejected. This means our final set of shapes will in all likelihood have fewer elements than ideally possible. I could not figure out a way to add this conditional to the test, at least not without scripting.

Attachments:

• Lshapes.gh, 21 KB
• LBlock.3dm, 25 KB

Permalink Reply by David Rutten on January 6, 2017 at 3:54am

This is using some C# to cull intersecting polylines. It gives a much tighter packing than my previous solution. It's also much faster.

Attachments:

• LShapes_CSharp.gh, 13 KB

Permalink Reply by Mohamed Naeim on January 5, 2017 at 3:29pm

I tried this 

 and this 

Attachments:

• move L shapes.gh, 33 KB

Permalink Reply by Mohamed Naeim on January 5, 2017 at 3:46pm

Because I did it in a hurry, I still got a crash when it exceed the number of possible fillings 

but it works like a charm if you can get the exact number of locations 

you will need anemone plugin

Attachments:

• move L shapes2.gh, 25 KB

Permalink Reply by Mohamed Naeim on January 6, 2017 at 5:25pm

try number 8 

I think this one works perfectly 

Attachments:

• move L shapes8.gh, 39 KB

Permalink Reply by Fth D on January 15, 2017 at 12:05pm

Thank you very very much! Both of the solutions are great! Final products are completely what I want. Mohammed, your solution is really attractive and totaly what I want. But the problem is your solution is so complicated to me. I can't understand and decode it. David's solution is more basic and understandable for a newbie. Thank you again again for you great work and help!