---
title: Hamiltonian Sheaf Overview
url: https://www.emergentmind.com/topics/hamiltonian-sheaf
type: topic
---

# Hamiltonian Sheaf Overview

A Hamiltonian sheaf is a mathematical construct in microlocal sheaf theory that provides a sheaf-theoretic quantization of Hamiltonian flows on cotangent bundles. This concept unifies derived category techniques, microlocal analysis, and symplectic dynamics. The theory centers on the existence and stability of certain sheaf kernels under Hamiltonian isotopies, with profound applications to invariants in symplectic topology, Lagrangian intersection theory, persistence modules, and categorical structures analogous to those in Fukaya–Floer theory. The Hamiltonian sheaf formalism, as developed by Guillermou–Kashiwara–Schapira, Tamarkin, and subsequent contributors, enables a translation of symplectic rigidity and energy estimates into microlocal or categorical terms.

## 1. Tamarkin Category and Hamiltonian Sheaves

The Tamarkin category $\mathcal{D}(X)$ is a triangulated subcategory of the (derived) category of sheaves on $X \times \mathbb{R}_t$ over a field $\mathbb{k}$, defined as the left orthogonal complement to objects microsupported in $\{\tau \le 0\} \subset T^*(X \times \mathbb{R}_t)$, where $\tau$ is the covector dual to the added “time” variable $t$. Explicitly,
\[
\mathcal{D}(X) := {}^\perp \mathbf{S}\mathcal{D}_{\tau \le 0}(\mathbb{k}_{X\times \mathbb{R}_t}),
\]
which can equivalently be realized as the essential image of the “positive-time convolution” projector $P_\ell(F) = \mathbb{k}_{X\times [0, \infty)} \star F$, killing microsupport in $\{\tau \le 0\}$ [2301.10598].

Elements of $\mathcal{D}(X)$ are interpreted as sheaves with a Hamiltonian time direction, admitting an exact autoequivalence via translation in $t$ (denoted $T^c$). The category becomes a natural home for quantizing Hamiltonian isotopies and analyzing their stability properties.

## 2. Sheaf Quantization of Hamiltonian Isotopies

Given a smooth manifold $M$ and a time-dependent Hamiltonian $H_s: T^*M \times I \to \mathbb{R}$ (with $I \supset [0,1]$) generating a Hamiltonian flow $\varphi_s^H$, the Guillermou–Kashiwara–Schapira (GKS) theory constructs a unique sheaf kernel
\[
\mathcal{K}^H \in \mathcal{D}(M \times M \times I)
\]
microsupported on the conic Lagrangian graph of the homogeneous lift of $\varphi_s^H$ in $T^*(M \times \mathbb{R}_t)$, normalized so $\mathcal{K}^H_0 = \mathbb{k}_{\Delta_M}$ [2502.19978, 1005.1517]. For each $s$, the sheaf $\mathcal{K}^H_s := \mathcal{K}^H|_{s}$ acts by convolution to transport sheaves’ microsupports under the flow.

The key GKS existence–uniqueness theorem asserts that for any homogeneous Hamiltonian isotopy on the punctured cotangent bundle, there exists a unique sheaf quantization kernel $K$ with $\operatorname{SS}(K)$ equal to the conic Lagrangian corresponding to the Hamiltonian flow, and $K$ restricts to the constant kernel on the diagonal at $t=0$ [2502.19978, 1005.1517, 2205.02661].

## 3. Interleaving Distance and Stability under Hamiltonian Flows

In $\mathcal{D}(X)$, the Tamarkin distance (also called the isomorphism distance $d_{\mathrm{isom}}$ or $\Delta$) measures how much “Hamiltonian energy” is required to deform one object into another. For $F,G \in \mathcal{D}(X)$, $(F,G)$ are said to be $(a,b)$–isomorphic if there exist morphisms
\[
\alpha: F \to T^a G,\quad \beta: G \to T^b F
\]
such that the composites recover canonical translations $T^{a+b}$ of $F$ and $G$. The Tamarkin distance is
\[
\Delta(F,G) := \inf\{a+b \mid (F,G)\ \text{are}\ (a,b)\text{–isomorphic}\} [2301.10598, 2201.02598].
\]

The main Hamiltonian-stability theorem states:
\[
\Delta(\mathcal{K}_1^0, \mathcal{K}_1^H) \le \|H\|_{\mathrm{osc}},
\]
where $\|H\|_{\mathrm{osc}} = \int_0^1 (\max_{p \in T^*M} H_s(p) - \min_{p \in T^*M} H_s(p))\, ds$ is the Hofer oscillation norm. This fundamental result shows that the Tamarkin metric in the sheaf category mirrors Hofer-type energy estimates from symplectic geometry [2301.10598].

With support constraints, stability can be refined: if $F \in \mathcal{D}_A(M)$ (microsupported over $A \subset T^*M$) and $H$ is a Hamiltonian, then for any continuous $f: I \to \mathbb{R}$,
\[
\Delta\left(F, T^{-c}(\mathcal{K}_1^H \star F)\right) \le B(H, f, A),
\]
where $B(H, f, A)$ bounds the energy over the evolved support of $A$ [2301.10598].

## 4. Sheaf-theoretic Invariants and Lagrangian Intersection Estimates

Sheaf quantization concretely encodes Lagrangian intersection theory. For a rational Lagrangian immersion $\iota: L \to T^*M$, Asano–Ike construct objects in the Tamarkin category whose self-Hom spaces recover the cohomology ring of $L$, with cup-products realized microsheaf-theoretically [2005.05088]. 

Using these constructions, one obtains explicit lower bounds for displacement energy:
\[
e(\iota, \varphi_1^H(\iota)) \geq \|H\|
\]
and intersection count estimates:
\[
\#(\iota(L) \cap \varphi_1^H(\iota(L))) \geq \sum_{i=0}^{\dim L} b_i(L)
\]
(where $b_i(L)$ are Betti numbers), and
\[
\#(\iota(L) \cap \varphi_1^H(\iota(L))) \geq cl(L) + 1
\]
for the cup-length $cl(L)$, under energy constraints [2005.05088]. These mirror classical Floer-theoretic inequalities but can be proven by microlocal sheaf methods, bypassing analytic issues such as transversality and $J$-holomorphic curves.

In the point case, the Tamarkin category reduces to the derived category of persistence modules, and $\Delta$ becomes the usual interleaving distance, yielding a categorical perspective on stability of Reeb graphs and sublevel set persistence [2301.10598].

## 5. Functorial Properties, Completeness, and Applications

The map $\varphi \mapsto \mathcal{K}^{\varphi}$ provides a functor from the group of Hamiltonian diffeomorphisms (under Hofer metric) to the Tamarkin category, and this assignment is Hofer-Lipschitz. Under suitable completeness hypotheses, this extends to homeomorphisms and non-smooth cases, as the Tamarkin category is complete with respect to the metric $d$ induced by interleaving [2201.02598]. For sequences $F_n$ with $\sum a_n < \infty$ and $d(F_n, F_{n+1}) \leq a_n$, there exists a limit $F_\infty$ [2201.02598].

The sheaf-theoretic spectrum and Lusternik–Schnirelmann invariants can also be formulated in this framework, yielding Arnold-type intersection theorems for Hamiltonian homeomorphisms [2201.02598].

## 6. Relations to Fukaya Categories and Floer Theory

The sheaf-theoretic approach provides a “Betti-model” for the Fukaya–Floer theory via the Nadler–Zaslow and Ganatra–Pardon–Shende quasi-equivalence:
\[
\mathrm{Sh}^w_\Lambda(X)^{op} \simeq \mathcal{W}(T^*X, \partial\Lambda)
\]
where $\mathrm{Sh}^w_\Lambda(X)$ is the category of wrapped constructible sheaves with microsupport in a conic Lagrangian stop $\Lambda$, and $\mathcal{W}$ is the wrapped Fukaya category [2205.02661]. In this context, Hamiltonian sheaves serve as categorical quantizations of Lagrangian branes and as functorial carriers for symplectic and topological invariants.

## 7. Extensions and Higher Structures

Recent work expands the Hamiltonian sheaf machinery to more general settings, including quantization via sheaves of spectra in symmetric monoidal $(\infty, 2)$-categories of correspondences, as in the context of Hamiltonian $\coprod_n BO(n)$-actions and their connections to Bott periodicity, the $J$-homomorphism, and stratified Morse theory [1902.06708]. Here, algebra and module structures arise from explicit correspondences in the category, and Morse-theoretic local systems become enriched by topological and categorical symmetry data.

Hamiltonian sheaf theory thus provides a categorical bridge between symplectic dynamics, microlocal analysis, and homotopical/categorical invariants, with broad ramifications in modern symplectic topology and representation theory.

Source: https://www.emergentmind.com/topics/hamiltonian-sheaf