Vuorinen survey Problems 3.38(a)–(b): Newton iteration for the inverse Grötzsch modulus

Let μ:(0,1)→(0,∞)\mu:(0,1)\to(0,\infty) be the Grötzsch modulus function μ(r)=π2K(1−r2)K(r)\mu(r)=\frac{\pi}{2}\frac{\mathcal{K}(\sqrt{1-r^2})}{\mathcal{K}(r)}, where K\mathcal{K} is the complete elliptic integral of the first kind. For y>π2y>\frac{\pi}{2}, set x0=1cosh⁡yx_0=\frac{1}{\cosh y} and define Newton's iteration for solving μ(x)=y\mu(x)=y by xn+1=xn−μ(xn)−yμ′(xn)x_{n+1}=x_n-\frac{\mu(x_n)-y}{\mu'(x_n)}. Does xnx_n converge to μ−1(y)\mu^{-1}(y) for every y>π2y>\frac{\pi}{2}? Moreover, for every y>πy>\pi, is the sequence strictly increasing, i.e. does xn<xn+1x_n<x_{n+1} hold for all n≥0n\ge 0?

References

Progress summary

Refreshed
Claimed solved

An unrefereed August 2026 preprint claims to settle both questions about this classical Newton method, but the result has not been independently verified.

Vuorinen’s survey problem asks for global convergence and monotonicity properties of Newton iteration for the inverse Grötzsch modulus.

August 2026 global-convergence claim

On August 24, 2026, a preprint claimed global convergence and monotonicity, together with a stronger monotonicity bound for y>πy>\pi; these results would resolve both identified questions.

Current status (as of August 2026): Both questions are claimed solved by an unrefereed preprint, but the claim remains unverified.

Sources

Solutions 0

No solutions have been posted yet.