2025年04月
The classical Gromov width measures the largest symplectic ball embeddable into a symplectic manifold; inspired by the symplectic camel problem, we generalize this to ask how large a symplectic ball can be embedded as a family over a parameter space $N$. Given a smooth map $f: N \to Ω$, where $Ω$ is a symplectic manifold, we define the \emph{parametric Gromov width} $\mathrm{Gr}(f,Ω)$ as the supremum of capacities $a>0$ for which there exists a family of balls, parametrized by $N$, of capacity $a$ whose centers trace out the map $f$. For Liouville domains $Ω$, we establish upper bounds on $\mathrm{Gr}(f,Ω)$ using the Floer cohomology persistence module associated to $Ω$. Specializing to fiberwise starshaped domains in the cotangent bundle $T^*M$, we derive computable bounds via filtered string topology. Specific examples of $Ω$ -- including disk cotangent bundles of thin ellipsoids, open books, and tori -- demonstrate our bounds, and reveal constraints on parameterized symplectic embeddings beyond the classical Gromov width.
@misc{parametric-gromov-width,
author = {Filip Broćić and Dylan Cant},
title = {Parametric {Gromov} width for {Liouville} domains},
year = {2025},
eprint = {2504.15207},
archivePrefix = {arXiv},
primaryClass = {math.SG},
}