3 ms·
I find these sort of calculation engines very interesting. Here is my attempt to solve the same somewhat more geometrically in solvespace. Any good papers on th
by somat 3y ago
I find these sort of calculation engines very interesting. Here is my attempt to solve the same somewhat more geometrically in solvespace. Any good papers on the data structures needed to implement one?
I am not very good at math so I might of done something terribly wrong but the theory is. The rectangles have equal area, area is length * width, items with the same L + W will have the same area(this is where I might have gone wrong), create a structure where all rectangles are forced to have the same L + W.
http://nl1.outband.net/extra/floorplan.png http://nl1.outband.net/extra/floorplan.png
http://nl1.outband.net/extra/floorplan.slvs http://nl1.outband.net/extra/floorplan.slvs
- PaulHoule 3y agoI could see this being useful for graphical layout.
- mk_stjames 3y agoThis is a class of problems known as constraint solvers, and a program like Solvespace is actually using a symbolic algebra system under the hood to 'solve' the layout as you constraint the sketch in the GUI in nearly the exact way that SymPy is solving the puzzle above. There are various approaches; Solvespace uses a modified Newton's method and was written in C++ making use of just the Eigen lib.
- WoodenChair 3y ago> items with the same L + W will have the same area(this is where I might have gone wrong), Unfortunately yeah that’s not right. Think of a 1x3 and a 2x2.
- somat 3y agoRight, a simple sanity check on the rects would have shown my mistake.