Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Gauss's Circle Problem [1] is to find the number of lattice points inside the boundary of a circle with a given radius and center at the origin. I'm interested in the "reversed" version of this problem. Given that I know the number of lattice points, what is the radius of the circle?

share|improve this question
3  
What is the exact statement of the problem you want to solve. How is it different than Gauss' problem? –  Boris Bukh Mar 31 '13 at 2:27

1 Answer 1

up vote 4 down vote accepted

The number of nodes for radius $r$ is $N_r \approx \pi r^2$ so you can invert to $r_N\approx\sqrt{\frac{N}{\pi}}$. If you wanted to derive the squared distance $r^2$ from $N$, you would have trouble but the square roots are bunched together so maybe that is accurate enough. This does leave some amount of uncertainty:

For $r$ from $\sqrt{1098}$ to $\sqrt{1105}-\varepsilon$ (SO $33.136$ to $33.24153$ and even a bit beyond that) we have $N_r=3456$. At $r=\sqrt{1105}=33.24154$ it jumps to $3488$ because there are $32$ points : $[4,33] ,[9,32],[12,31]$ and $[23,24]$ along with all the variations of making one or both negative and/or swapping the order.

So for $N=3456$ we have $\sqrt{\frac{N}{\pi}}=33.167=\sqrt{1100}$, a bit low perhaps. For $N=3488$ we have $\sqrt{\frac{N}{\pi}}=33.3206 =\sqrt{1110.26}$ This is high because $N$ jumps from $3488$ to $3496$ at $r=\sqrt{1108}$ and then jumps again to $3504$ at $r=\sqrt{1109}$

If you use known precise error bounds for the usual problem you could get error bounds for this dual problem but you can't avoid the uncertainty demonstrated above using a uniform formula.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.