The equation has infinitely many solution for any coprime integers and . And every solution is given by the translate of a fixed solution by . One can find the solution by the gcd Euclidean algorithm applied to and . (Thus a solution can be obtained by looking at the last convergent in the continued fraction expansion of .
Question: How does a “minimal” solution look? If are of some size . what can we say about the relative sizes or ? How are they are distributed?
We are essentially asking for how skewed is the fundamental parallelogram of the integer lattice with a given basis vector .
The answer is for random of some size, the second basis vector looks completely random– that is the skewness is uniformly distributed.
An analytic way to prove this is using the fact that is the inverse , and hence the uniform distribution reduces to cancellation in Kloosterman sums.
See: Akio Fujii. On a problem of Dinaburg and Sina ̆ı. Proc. Japan Acad. Ser. A Math. Sci., 68(7):198–203, 1992
Because the solution are related to the Euclidean algorithm and continued fractions, one can study the problem using the distributions/dynamics of continued fractions and Gauss map. This gives uniform distribution at a smaller scales than the Kloosterman sums can establish.
See Dimitry Dolgopyat. On the distribution of the minimal solution of a linear Diophantine equation with random coefficients.Funktsional. Anal. i Prilozhen., 28(3):22–34, 95, 1994
Lattice and Modular surface- We can also see that the produce reduces to the uniform distribution of modulo one as in the modular group varies. To capture this analytically using to Weyl’s criterion, we need to study sums over closely related to the Poincare series, and establish their analytic properties.