Convex-hull conjecture for Orthomin(1)

About 15 years old · traced to

Let ζ1,…,ζd∈C\zeta_1,\dots,\zeta_d\in\mathbb{C} be the spectral points used in the finite-dimensional Orthomin analysis, and let Hull⁡(ζ1,…,ζd)\operatorname{Hull}(\zeta_1,\dots,\zeta_d) denote their convex hull. Let qnq_n be the Orthomin(1)(1) residual convergence factor and let ρ\rho be the parameter in the associated linear system. Convex-hull conjecture. If

−ρ∈Hull⁡(ζ1,…,ζd),-\rho\in\operatorname{Hull}(\zeta_1,\dots,\zeta_d),

then

qn→ρ.q_n\to\rho.

The paper describes this as a sharpened version of its earlier conjecture. The preceding proposition proves only the complementary obstruction that −ρ-\rho outside the convex hull implies qnq_n does not converge to ρ\rho; the stated implication is therefore left open in the supplied text.

References

Primary source

Andrei Draganescu and Florin Spinu, “Sharp estimates for the convergence rate of Orthomin(k) for a class of linear systems”, arXiv:1109.2669 (2025).

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.