Limit-point conjecture for the generalized Rogers-Ramanujan continued fraction

Let a=ζka=\zeta_k and x=ζmx=\zeta_m be primitive roots of unity, with m3m\geq3. Limit-point conjecture. The continued fraction Ka(x)K_a(x) converges exactly when

14+amQ(a,x)and14+amR0\sqrt{\frac14+a^m}\notin\mathbb Q(a,x)\quad\text{and}\quad \frac14+a^m\notin\mathbb R_{\leq0}

and hence exactly when 5m5\nmid m or kmk\nmid m, and am1a^m\neq-1. In the convergent case its limit is the one given by the paper's general limit formula. If 5m5\mid m and kmk\mid m, the two limit points are

Pm2(a,x)12±14+amaxm1Pm3(a,x).\frac{P_{m-2}(a,x)}{\frac12\pm\sqrt{\frac14+a^m}-ax^{m-1}P_{m-3}(a,x)}.

For 3m3\mid m and am=1a^m=-1, these limit points may also occur; otherwise, the remaining subsequences produce exactly the three additional limit points specified by the source. This conjectural picture organizes convergence, divergence, and all predicted limit points of the continued fraction at primitive roots of unity.

Sources & referencesView supporting material

Primary source

Emil-Alexandru Ciolan and Robert Axel Neiss, “Convergence properties of the classical and generalized Rogers-Ramanujan continued fraction”, arXiv:1504.06482 (2015).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.