Answers are grounded in this paper and cite their sources.
This paper continues our earlier work, in which the differential evolution algorithm was applied to coefficient problems in Geometric Function Theory, and the sharp 1/8 bound of the fourth Hankel determinant detH4,1(f) was identified numerically in the class S* of starlike functions, i.e., of analytic univalent functions f in the unit disk D normalized by f(0)=0 and f′(0)=1, for which f(D) is starlike with respect to the origin. Here, the approach is extended to the fifth and sixth Hankel determinants detH5,1(f) and detH6,1(f). Using finite-dimensional reduction in the extremal problem provided by the Hummel–Sakaguchi theorem on support points of the Carathéodory class, we perform a fully documented and reproducible differential evolution search. In all independent runs, the values found agree, to within 10−14, with 16/625 and 1/243, respectively, and all maximizers found coincide, up to rotation, with the q-fold symmetric Koebe function fq(z)=z(1−zq)−2/q, q=5,6. We prove that |detHq,1(fq)|=(2/q)q−1 for every q∈N so that these values are attained in S*. Combining the numerical computations with the known sharp bounds for q≤3, we conjecture that |detHq,1(f)|≤(2/q)q−1 for all f∈S* and q=4,5,6, with equality for fq and its rotations. In the whole class S*, these upper bounds are supported by numerical computations but are not proved; they are stated explicitly as conjectures. We also show that the pattern does not extend to all orders: For every q≥10, the function fq−1 satisfies |detHq,1(fq−1)|=2q−2(q−3)/(q−1)q>(2/q)q−1 so that fq is not extremal for q≥10 and the general inequality |detHq,1(f)|≤(2/q)q−1, q∈N, is false.
No takes yet. Share an insight, caveat, or question.
Jastrzębski et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: