Brownian motion and stochastic areas on complex full flag manifolds

Fabrice Baudoin, Nizar Demni, Teije Kuijper, Jing Wang · 2025

Read Brownian motion and stochastic areas on complex full flag manifolds by Fabrice Baudoin, Nizar Demni, Teije Kuijper, Jing Wang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-09124.txt.

Introduction

Fabrice Baudoin, Nizar Demni, Teije Kuijper, and Jing Wang study random motion on a curved space of flags, in a math paper from 2025. A flag here is a nested set of planes, and random motion on it is hard to track in a direct way. The team seeks exact laws for areas swept out along the way.

They show the motion can be built from a simpler random walk on matrices. That link yields a clear formula for the joint law of the swept areas. In the long run, those areas follow a known heavy-tailed law with separate parts. The work also gives new results on winding around spheres.

Catalog record

CREATORFabrice Baudoin, Nizar Demni, Teije Kuijper, Jing Wang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE74 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-09124

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Brownian motion and stochastic areas on complex full flag manifolds Fabrice Baudoin, Nizar Demni, Teije Kuijper, Jing Wang Abstract We show that the Brownian motion on the complex full flag manifold can be represented by a matrix-valued diffusion obtained from the unitary Brownian motion. This representation actually leads to an explicit formula for the characteristic function of the joint distribution of the stochastic areas on the full flag manifold. The limit law for those stochastic areas is shown to be a multivariate Cauchy distribution with independent and identically distributed entries. Using a deep connection between area functionals on the flag manifold and winding functionals on complex spheres, we establish new results about simultaneous Brownian windings on the complex sphere and their asymptotics. As a byproduct, our work also unveils a new probabilistic interpretation of the Jacobi operators and polynomials on simplices. 1 Introduction The study of stochastic area functionals has deep roots in both probability and geometry, tracing back to Paul Lévy’s foundational work [ 15 ] on planar Brownian motion. Over the years, this subject has evolved into a rich and extensive theory; see [ 5 ] for a recent survey. One of the main motivations of this paper is to extend this theory to the setting of flag manifolds. Flag manifolds play key roles in differential geometry [ 3 ] , representation theory [ 2 ] , algebraic geometry [ 10 ] , physics [ 17 ] , and numerical analysis [ 19 ] . Among them, the complex full flag manifold F_{1,2,\ldots,n-1}(\mathbb{C}^{n}) parametrizes nested sequences of complex subspaces: \{0\}\subsetneq W_{1}\subsetneq\cdots\subsetneq W_{n-1}\subsetneq\mathbb{C}^{n}. The importance of this space stems from the fact that any complex partial flag manifold can be obtained from it via a submersion [ 10 , §1.2] . It also admits a homogeneous Riemannian structure as \mathbf{U}(n)/\mathbf{U}(1)^{n} , where \mathbf{U}(n) is the unitary group and \mathbf{U}(1)^{n} its maximal torus of diagonal unitary matrices. Though not a symmetric space, the full flag manifold has a complex Kähler structure [ 9 ] , which will be a crucial ingredient in our investigations of: the Brownian motion on F_{1,2,\ldots,n-1}(\mathbb{C}^{n}) ; its associated stochastic area functionals; and its connections to Brownian winding on complex spheres. Brownian motion on Lie groups and homogeneous spaces has long been a cornerstone of stochastic differential geometry [ 4 , 13 , 16 ] . A key contribution of this paper is the construction of Brownian motion on the full flag manifold via projection from unitary Brownian motion. Expressing this process in local affine coordinates derived from the quotient structure allows us to explicitly compute its generator, which governs the radial dynamics of the process. In particular, these dynamics will be identified with Jacobi diffusions on simplices. Building on the general theory described in [ 5 ] , we then use the Kähler structure of \mathbf{U}(n)/\mathbf{U}(1)^{n} to define a natural stochastic area process. Specifically, the Riemannian fibration \mathbf{U}(1)^{n}\to\mathbf{U}(n)\to\mathbf{U}(n)/\mathbf{U}(1)^{n} allows us to view \mathbf{U}(n) as a torus bundle over the full flag manifold, leading to an intrinsic n -dimensional area process. For a horizontal Brownian motion on \mathbf{U}(n) , these stochastic area measures accumulated phase differences across the torus fibers. Using a skew-product decomposition, we derive the joint characteristic function of these areas and establish that, as t\to+\infty , their limiting distribution follows a multivariate Cauchy law with independent components after proper rescaling. This limit theorem relies on spectral properties of Jacobi operators and their associated orthogonal polynomials on simplices. A striking application arises in the study of simultaneous Brownian windings on the complex sphere \mathbb{S}^{2n-1}\subset\mathbb{C}^{n} , generalizing a result in [ 6 ] . By linking the stochastic areas of the flag manifold to angular windings of spherical Brownian motion, we establish asymptotic independence of the winding processes. Specifically, upon proper rescaling, these windings also converge in distribution to independent Cauchy random variables. The paper is organized as follows: Section 2 reviews complex flag manifolds and Jacobi polynomials on simplices; Section 3 constructs the Brownian motion on the full flag manifold as a suitable projection of the unitary Brownian motion and analyzes its radial dynamics; Section 4 introduces stochastic area functionals, derives the corresponding characteristic functions, and contains a proof of convergence of said functionals to a multivariate Cauchy distribution; and finally, Section 5 applies these results to simultaneous Brownian windings on complex spheres, establishing asymptotic laws and connections to Euclidean Brownian motion. 2 Preliminaries This section is devoted to recalling some geometric and analytic preliminaries that will be used in the subsequent analysis. We begin by reviewing basic geometric facts about the flag manifold and the Riemannian submersion structure it carries. We then turn to a brief overview of the Jacobi operator and Jacobi polynomials on simplices, which will play a key role in the later sections. Throughout the paper, let n\geq 2 be an integer. 2.1 Complex full flag manifolds In this section, we review the geometric structure of complex full flag manifolds, including their realization as homogeneous spaces and the associated Riemannian submersion structure. Definition 2.1 . A flag (W_{1},\dots,W_{k}) of \mathbb{C}^{n} is a sequence \displaystyle\{0\}=:W_{0}\subsetneq W_{1}\subsetneq\dots\subsetneq W_{k}\subsetneq W_{k+1}:=\mathbb{C}^{n} of complex subspaces of \mathbb{C}^{n} . The number k is called the length of the flag and the k -tuple \displaystyle(\dim(W_{1}),\dots,\dim(W_{k})) the signature of the flag. Definition 2.2 . Let V be a complex vector space of dimension n . The (complex) partial flag manifold F_{d_{1},\dots,d_{k}}(\mathbb{C}^{n}) of signature (d_{1},\dots,d_{k}) is the collection of flags of \mathbb{C}^{n} of signature (d_{1},\dots,d_{k}) . The (complex) full flag manifold is the flag manifold F_{1,2,\dots,n-1}(\mathbb{C}^{n}) . To study and define the smooth structure on F_{1,2,\dots,n-1}(\mathbb{C}^{n}) we will make a first identification. Note that the subgroup of invertible upper triangular matrices T^{n}(\mathbb{C}) acts transitively on the set \mathrm{GL}_{n}(\mathbb{C}) of n\times n invertible matrices by right multiplication and that we can identify F_{1,2,\dots,n-1}(\mathbb{C}^{n}) as the quotient space \mathrm{GL}_{n}(\mathbb{C})/T^{n}(\mathbb{C}) . Indeed, consider the canonical basis e_{1},\dots,e_{n} of \mathbb{C}^{n} , then the surjective map \displaystyle\begin{array}[]{ccc}\mathrm{GL}_{n}(\mathbb{C})&\to&F_{1,2,\dots,n-1}(\mathbb{C}^{n})\\ M&\mapsto&(W_{1},\dots,W_{n-1})\end{array} with W_{i}=\mathrm{span}(Me_{1},Me_{2},\dots,Me_{i}) is invariant by this action and thus descends into a bijection \mathrm{GL}_{n}(\mathbb{C})/T^{n}(\mathbb{C})\to F_{1,2,\dots,n-1}(\mathbb{C}^{n}) . Since T^{n}(\mathbb{C}) acts smoothly and properly on \mathrm{GL}_{n}(\mathbb{C}) by multiplication from the right, one has the identification \displaystyle\mathrm{GL}_{n}(\mathbb{C})/T^{n}(\mathbb{C})\cong F_{1,2,\dots,n-1}(\mathbb{C}^{n}), and the smooth structure is the unique one turning the canonical projection into a smooth submersion. In particular, the complex dimension of F_{1,2,\dots,n-1}(\mathbb{C}^{n}) is given by: n^{2}-\frac{n(n+1)}{2}=\frac{n(n-1)}{2}. In order to define a Riemannian metric on F_{1,2,\dots,n-1}(\mathbb{C}^{n}) , we appeal to its compact realization: \displaystyle F_{1,2,\dots,n-1}(\mathbb{C}^{n})\cong\mathbf{U}(n)/\mathbf{U}(1)^{n}, where \mathbf{U}(n) is the unitary group : \displaystyle\mathbf{U}(n):=\{M\in\mathbb{C}^{n\times n}\mid M^{*}M=I_{n}\}. This realization is actually obtained in a similar fashion as the one above and the action is given by by right multiplication from \mathbf{U}(1)^{n} , identified with the set of diagonal matrices in \mathbf{U}(n) . The Riemannian metric on F_{1,2,\dots,n-1}(\mathbb{C}^{n}) is then the unique one making the canonical projection \pi:\mathbf{U}(n)\to\mathbf{U}(n)/\mathbf{U}(1)^{n} a Riemannian submersion, where \mathbf{U}(n) is equipped with its bi-invariant metric induced by the Killing form. This Riemannian submersion \pi has totally geodesic fibers isometric to \mathbf{U}(1)^{n} , since it is a Bérard-Bergery fibration [ 8 , Theorem 9.80] . One can also see F_{1,2,\dots,n-1}(\mathbb{C}^{n}) as an algebraic sub-variety of \mathbb{C}P^{n-1}\times\dots\times\mathbb{C}P^{n-1} , where \mathbb{C}P^{n-1} is the complex projective space, i.e. the set of complex lines in \mathbb{C}^{n} . Indeed, one has the embedding of F_{1,2,\dots,n-1}(\mathbb{C}^{n}) into

Read the complete text