Berline–Vergne localization formula

Berline–Vergne localization formula is a key result in equivariant cohomology and index theory that expresses integrals of equivariantly closed forms over compact manifolds with torus actions as sums localized at fixed points. The formula links techniques from symplectic geometry, representation theory, and differential topology, and it influenced developments in mathematical physics, index theory, and combinatorics.

Statement

The formula applies to a compact smooth manifold M equipped with an action of a compact torus T and an equivariant cohomology class α in H_T^*(M). Under nondegeneracy assumptions on the infinitesimal action at fixed points, the integral ∫_M α can be computed as a finite sum over the connected components F of the fixed point set M^T, each term involving the restriction α|_F and the equivariant Euler class e_T(N_F) of the normal bundle N_F. Berline and Vergne gave a precise analytic expression using equivariant differential forms and equivariant Thom forms, providing a localization identity that parallels the earlier work of Atiyah and Bott on Lefschetz-type formulas and the Duistermaat–Heckman formula in symplectic geometry.

Historical context and development

The formula emerged in the early 1980s following interactions among several streams of research: index theory, symplectic geometry, and mathematical physics. The work of Michael Atiyah, Raoul Bott, and Friedrich Hirzebruch on fixed-point theorems and index formulas provided conceptual groundwork; contemporaneous contributions by Victor Guillemin, Shlomo Sternberg, and Elias Stein in symplectic and representation theory shaped applications. Nicole Berline and Michèle Vergne synthesized analytic localization techniques influenced by Alexander Kirillov's orbit method and Daniel Quillen's superconnection formalism, while the Duistermaat–Heckman theorem and the Atiyah–Singer index theorem framed broader implications. Subsequent developments involved interactions with Edward Witten's work on supersymmetric quantum mechanics and topological field theory, and with contributions by Maxim Kontsevich and Mikhail Gromov in mirror symmetry and enumerative geometry.

Proof outline and methods

Berline and Vergne constructed an analytic proof using equivariant differential forms, heat kernel methods, and localized Thom forms. The argument begins by choosing an invariant Riemannian metric and an equivariant connection to define an equivariant curvature; then one constructs an equivariant form whose equivariant differential equals a given closed form except on a tubular neighborhood of the fixed locus. By applying stationary phase and Gaussian integral estimates, contributions concentrate near fixed components, and one computes each local contribution via equivariant normal form coordinates and Gaussian integrals. Techniques parallel those in the proofs of the Atiyah–Singer index theorem by Patodi and Nicole Berline herself, and use tools related to Quillen superconnections, Getzler rescaling, and heat kernel asymptotics. Alternative proofs rely on localization in equivariant cohomology as in the work of Atiyah and Bott, or on algebraic localization formulas used by Jean-Michel Bismut in analytic torsion and index theory.

Applications and examples

The localization formula has been applied across several areas. In symplectic geometry it is instrumental for the Duistermaat–Heckman measure computation for Hamiltonian actions studied by Guillemin and Sternberg, and appears in computations of volumes and intersection pairings on moduli spaces such as the moduli space of flat connections investigated by Nigel Hitchin and Michael Thaddeus. In representation theory it underlies character formulas related to the orbit method of Kirillov and to the work of Bertram Kostant. In equivariant index computations it complements the Atiyah–Bott fixed-point theorem and the Atiyah–Singer index theorem, and it is used in localization techniques for partition functions in Edward Witten's supersymmetric gauge theories and in the analysis of Chern–Simons invariants studied by Witten and Graeme Segal. Specific examples include localization on toric manifolds as in the theory of Victor Guillemin and Mark Gross, computations on flag varieties connected to the work of Alexander Beilinson and Joseph Bernstein, and exact evaluations of integrals on Grassmannians related to William Fulton and Robert MacPherson.

Generalizations and related results

Generalizations extend to nonabelian group actions via abelianization techniques credited to Atiyah, Bott, and Harish-Chandra, and to equivariant K-theory with versions by Michel Atiyah and Graeme Segal. Extensions to orbifolds and singular spaces connect with the Kawasaki index theorem and the work of Boris Tsirelson. Formulations in algebraic geometry using localization in equivariant Chow groups were developed by William Fulton and Ravi Vakil, while deformation quantization perspectives relate to Kontsevich formality and Maxim Kontsevich’s work on Poisson manifolds. Connections to mirror symmetry, Gromov–Witten invariants by Mikhail Gromov and Rahul Pandharipande, and applications in combinatorics via Jeffrey–Kirwan residue formulas and the work of Lisa Jeffrey and Frances Kirwan further expanded the reach of localization techniques. Contemporary research links the formula to derived algebraic geometry by Jacob Lurie and to equivariant elliptic cohomology studied by Haynes Miller and Matthew Ando.

Category:Equivariant cohomology Category:Index theory Category:Symplectic geometry