Equivariant Localization in Geometry & Physics

• Equivariant localization is a set of techniques that reduce integrals and indices on spaces with group actions to computations on fixed-point or critical loci.
• It extends the classical ABBV formula to K-theory, non-abelian, and categorical settings, enabling explicit evaluations in geometry, topology, and quantum field theory.
• Applications range from computing characteristic classes in symplectic geometry to evaluating partition functions in supersymmetric theories and supergravity.

Equivariant localization is a principle and a suite of techniques in geometry, topology, representation theory, and mathematical physics for reducing integrals, indices, and invariants on spaces with group actions to computations on distinguished subspaces—the fixed-point loci or more generally, loci specified by critical points of auxiliary functions—organized according to the symmetries of the action. Foundational results include the Atiyah–Bott–Berline–Vergne (ABBV) theorem for equivariant cohomology, non-abelian, K-theoretic, and categorical generalizations, as well as extensions to quantum field theory, algebraic stacks, and derived geometry. The central idea is that, under suitable conditions, equivariant characteristic classes, partition functions, or indices localize to explicit sums or integrals over fixed-point sets, with contributions weighted by data from the normal bundle or category-theoretic invariants.

1. Classical Equivariant Localization: The ABBV Formula

Let $G$ be a compact Lie group acting smoothly on a compact, oriented manifold $M$. The foundation of equivariant localization is the construction of equivariant cohomology $H_G^*(M)$, for which the Cartan model $d_G = d - \iota_\xi$ (with $\xi$ running over the Lie algebra $\mathfrak{g}$ of $G$) provides a differential complex. Equivariant characteristic classes and integration are defined on this model.

When $T$ is a torus acting with isolated fixed points $\{p\}$, the ABBV localization theorem (Tu, 2013, Notman et al., 2023) states that for any equivariantly closed form $\alpha_T \in H_T^{2n}(M)$,

$M$0

where $M$1 is the equivariant Euler class of the tangent space at $M$2, computable from the torus weights on $M$3.

For more general fixed-point sets $M$4 (possibly higher-dimensional), the formula incorporates integration over $M$5 and division by the equivariant Euler class $M$6 of the normal bundle: $M$7 This formula has direct computational applications in characteristic numbers, Gromov–Witten invariants, and volume integrals in symplectic geometry (Notman et al., 2023, Tu, 2013).

2. K-Theory, Operational K-Theory, and GKM Descriptions

Localization in equivariant K-theory and related operational theories generalizes the ABBV result. For a torus $M$8 acting on a variety $M$9, Anderson–Payne's operational $H_G^*(M)$0-theory ring $H_G^*(M)$1 admits a localization theorem of Borel–Atiyah–Segal type: the restriction $H_G^*(M)$2 becomes an isomorphism after localizing at the multiplicative set generated by nontrivial characters (Gonzales, 2014).

In the case of T-skeletal varieties (finite fixed-point set and finitely many invariant curves), $H_G^*(M)$3 admits a Goresky–Kottwitz–MacPherson (GKM) description: classes correspond to collections $H_G^*(M)$4 such that $H_G^*(M)$5 for every invariant curve $H_G^*(M)$6 (with character $H_G^*(M)$7) connecting $H_G^*(M)$8 and $H_G^*(M)$9 (Gonzales, 2014).

In equivariant cohomology, rational rings of varieties such as real Grassmannians are determined by their GKM graphs, with additive and multiplicative structures determined by fixed-point data, torus weights, and edge congruences (He, 2016).

3. Non-Abelian and Categorical Localization

Equivariant localization extends beyond abelian symmetry and classical cohomology through non-abelian and categorical frameworks. When $d_G = d - \iota_\xi$0 is an algebraic (derived) stack equipped with a $d_G = d - \iota_\xi$1-stratification $d_G = d - \iota_\xi$2 (in the sense of Harder–Narasimhan theory or GIT), the non-abelian localization theorem relates $d_G = d - \iota_\xi$3-theoretic invariants on $d_G = d - \iota_\xi$4 to those on the centers $d_G = d - \iota_\xi$5 of the strata.

The virtual $d_G = d - \iota_\xi$6-theoretic formula (Halpern-Leistner, 28 Sep 2025) at the level of $d_G = d - \iota_\xi$7-groups is: $d_G = d - \iota_\xi$8 where $d_G = d - \iota_\xi$9 is the virtual Euler class of the normal complex, and $\xi$0 is the sharp pullback defined via the Thom isomorphism and baric weights. This is refined by a categorical decomposition: the category of highest-weight $\xi$1-homology cycles on $\xi$2 splits as the direct sum over cycles on the centers,

$\xi$3

This framework provides a universal wall-crossing formula for the difference of indices across variation of stability conditions and yields finiteness theorems for the cohomology of tautological complexes on moduli stacks (Halpern-Leistner, 28 Sep 2025).

4. Equivariant Localization in Quantum Field Theory and Supergravity

Equivariant localization principles are fundamental in exact computations of partition functions, indices, and observables in supersymmetric and topological field theories. In the Batalin–Vilkovisky (BV) framework (Cattaneo et al., 28 Jan 2025), localization arises from nilpotent differentials $\xi$4 built from symmetry generators, with integrals reducing via stationary phase or fixed-point theorems to sums over critical loci.

In supergravity, the presence of a supersymmetric background with an $\xi$5-symmetry Killing vector $\xi$6 allows the construction of equivariantly closed polyforms—combinations of spinor bilinears and background fields annihilated by $\xi$7 (Genolini et al., 2023, Genolini et al., 2024, Genolini et al., 11 Aug 2025, Genolini et al., 9 Apr 2026). The integral of such forms computes on-shell observables (free energies, central charges, black hole entropies, etc.) via the Berline–Vergne–Atiyah–Bott (BVAB) fixed-point formula: $\xi$8 This reduces the computation to fixed-point data: values of background fields, Chern classes, and normal bundle weights at the fixed loci. This principle applies in $\xi$9 and $\mathfrak{g}$0 gauged supergravity, including higher-derivative theories, and is robust under inclusion of nontrivial moduli or arbitrary higher-derivative couplings (Genolini et al., 9 Apr 2026, Genolini et al., 2024, Genolini et al., 11 Aug 2025).

Applications include explicit evaluation of supergravity partition functions, derivation of universal wall-crossing formulas for indices, and complete reduction of certain gravitational path integrals to combinatorial data from the symmetry action, without need for explicit solution of PDEs (Genolini et al., 2023, Genolini et al., 2023, Genolini et al., 11 Aug 2025).

5. Generalizations: Varying Polarization and Degenerate/Non-Compact Cases

Equivariant localization principles have been extended to non-compact, possibly singular, or non-symplectic settings using the notion of polarization or taming maps (Harada et al., 2010). For a Hamiltonian $\mathfrak{g}$1-manifold $\mathfrak{g}$2, a taming map $\mathfrak{g}$3 defines a vector field $\mathfrak{g}$4. The localization formula then decomposes global invariants into contributions from the localizing set $\mathfrak{g}$5, with each component modeled by a $\mathfrak{g}$6-polarized neighborhood.

This framework interpolates between classical ABBV localization (constant $\mathfrak{g}$7; localization to fixed points), norm-square localization (Witten–Paradan–Woodward; $\mathfrak{g}$8) and polytope decompositions (e.g., Brianchon–Gram). Cobordism invariance and the use of non-compact completions are essential in these generalizations (Harada et al., 2010).

6. Factorization Homology, Quantum Field Theory, and Quantization

Equivariant localization structures the computation of factorization homology in equivariant quantum field theory settings. For varieties $\mathfrak{g}$9 with $G$0 action and $G$1 a $G$2-equivariant factorization algebra, the localization theorem asserts that the natural map $G$3 becomes an isomorphism after inverting the Euler classes associated to nontrivial weights (Butson, 2020). This underpins the construction of QFT invariants, quantizations, and relationships between symplectic and holomorphic-twisted structures (e.g., $G$4-background, chiral algebras).

7. Chromatic and Bousfield Localizations in Equivariant Homotopy Theory

In stable homotopy theory, equivariant localization includes Bousfield localization with respect to thick subcategories of $G$5-spectra. The preservation of ring or $G$6-algebra structures under equivariant localization depends on compatibility with norm functors across conjugacy classes. Notably, localization at a nonequivariant spectrum (inflated to a $G$7-spectrum) always preserves genuine equivariant commutative ring structures (Hill, 2017).

Theoretical classification of acyclic categories, fracture squares, and cofiber sequences underpins the global structure of equivariant stable homotopy categories and their localization theorems.

---

Equivariant localization thus provides a versatile and unifying principle, linking geometric fixed-point theory, representation-theoretic decompositions, index theory, derived algebraic geometry, and quantum field theoretic computations. The explicit formulas, categorical decompositions, and wall-crossing structures furnish both profound conceptual insight and practical computational tools across modern geometry and mathematical physics.