Almost diophantine equations?

63 Views Asked by At

Given square tiles of dimension 4.27" and 1.87" plus or minus .01" what grout widths are possible to make low order least common multiples?

Grout width can vary by 10% without being seen as different.

The idea is that N tiles of tile X + G (N-1) = M tiles of tile Y + G (M-1)

So that in a running pattern at least in one dimension NX tiles can be replaced by MY tiles. (There is always 1 less groutline than tiles.)

The goal is to find solutions where for a given pair of tile sizes X and Y, and a small range of G that M and N are small numbers.

This is not strictly speaking a diophantine equation, except that it is close to a diophantine equation in nature in that the desired solution is limited to integers, even though all the constraints are rationals.

I've been playing around it with a spreadsheet.

In general I've found less bad solutions if the tiles have a larger common factor f. I'm bastardizing common factor here. D = i * f where D is the dimension of the tile, and i is an integer. f can be floating. E.g. A 4.25" tile and a 1.75" tile have a common factor of 0.25, since both tiles are integer multiples of this. I've also found less bad solutions when the grout width is either a divisor or a multiple of this common factor.

Is there a name for this class of problem?

What keywords do I need to look up to find better algorithms for finding solutions?

Warning: Not a mathematician. Be gentle. I'm guessing at tags. Feel free to adjust.