Cusp exponent for the amplitude-augmented functional

Determine how adding the amplitude term \(\lambda_3\int u^p/p\) to the functional \(\int_0^{e_0}(\lambda_1|u'|^2/2+\lambda_2/u)\,e\) affects the cusp exponent of the minimizing dissipation schedule.

Background

The paper completely solves the regularized least-time problem with Dirichlet roughness penalty and reciprocal-time cost, obtaining a unique cycloidal minimizer whose behavior near zero risk is u∗(e)∼(9/(2λ))1/3e2/3u^*(e)\sim (9/(2\lambda))^{1/3}e^{2/3}. It also discusses a more general functional from prior work that includes an additional nonnegative amplitude term λ3∫up/p\lambda_3\int u^p/p.

Although the existence argument is stated to extend to this amplitude-augmented functional, the authors do not obtain a closed-form identification of the minimizer or determine whether the additional term changes the near-zero cusp exponent. Establishing that effect is therefore explicitly left unresolved.

References

The amplitude term \lambda_3\int u{p}/p of is not covered by Theorem~\ref{thm:main}: the existence argument of Section~\ref{sub:existence} goes through unchanged, since that term is nonnegative and passes to the limit under uniform convergence, but the identification step has no closed form, and what the amplitude term does to the cusp exponent we leave open.

— Least-time Gradient Flow  (2610.01426 - Betti et al., 1 Oct 2026) in Remark on scaling (Remark \ref{rem:scaling})