Complex Winding Invariants: Theory & Applications

• Complex winding invariants are topological quantities that extend classical winding numbers, allowing non-integer and fractional values in settings with singularities or self-intersections.
• They are defined using methods such as Cauchy principal values, real-algebraic formulations, and polynomial invariants, enabling robust analysis in both continuous and discrete systems.
• These invariants play a critical role in applications ranging from non-Hermitian quantum systems and Floquet topologies to combinatorial graph theory and stability analysis.

Complex winding invariants are topological quantities that generalize the classical notion of the winding number for complex-valued functions and closed curves, with crucial roles in algebraic topology, complex analysis, mathematical physics, non-Hermitian topology, and combinatorial graph theory. While the integer winding number measures the signed number of times a closed curve encircles a point or a singularity, complex winding invariants can be non-integer, encode more sophisticated geometric and algebraic information, and remain meaningful or computable in higher complexity settings: for example, when the point of interest lies on the curve, in non-Hermitian quantum systems, for discrete color-exchange models, or for almost embeddings of graphs in the plane. Their definitions can employ Cauchy principal values, Green's function topology, real-space traces, or polynomial invariants of winding vectors, depending on context.

1. Classical and Generalized Complex Winding Numbers

The classical complex winding number for a closed, piecewise $C^1$ curve $\gamma\subset\mathbb{C}$ with basepoint $z_0\notin\gamma$ is

$$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$

a homotopy invariant, integer-valued, and equal to the degree of the map $t\mapsto (\gamma(t)-z_0)/|\gamma(t)-z_0|$ from the parameter domain to $S^1$. This measures how many times $\gamma$ winds around $z_0$ (Hungerbühler et al., 2018).

In the presence of singularities or when $z_0\in\gamma$, this integral diverges. Hungerbühler and Wasem introduce a Cauchy principal value generalization: $$w_{\mathrm{gen}}(\gamma, z_0)=\mathrm{PV}~\frac{1}{2\pi i}\int_\gamma \frac{dz}{z-z_0},$$ which can be equivalently written as a real integral with bounded integrand: $\gamma\subset\mathbb{C}$0 with $\gamma\subset\mathbb{C}$1. This yields non-integer windings, including fractional values (e.g., half-integers) at corner points or self-intersections. Crucially, $\gamma\subset\mathbb{C}$2 is invariant under deformations preserving the local angle at $\gamma\subset\mathbb{C}$3, and for a corner angle $\gamma\subset\mathbb{C}$4 at $\gamma\subset\mathbb{C}$5, $\gamma\subset\mathbb{C}$6. This allows formulation of a generalized residue theorem applicable when poles lie on the integration path, with the principal value of the contour integral picking up the generalized winding weights of residues (Hungerbühler et al., 2018).

2. Algebraic and Computational Approaches

In constructive and computational settings, winding numbers are re-expressed in terms of real-algebraic invariants. Li and Paulson provide a formalization in Isabelle/HOL, encoding curves as parameterized maps and winding numbers as Cauchy integrals. The central theorem relates the winding number to the Cauchy index of the imaginary-to-real part ratio of the curve relative to $\gamma\subset\mathbb{C}$7: $\gamma\subset\mathbb{C}$8 This can be decided algorithmically via sign changes of explicit polynomial functions, reducing analytic root-counting (via the argument principle) to Sturm-Tarski and variation sequences for polynomials in rectangles or half-planes (Li et al., 2018).

3. Complex Winding Invariants in Non-Hermitian and Floquet Topology

The theory of complex winding invariants has been pivotal in the topological classification of non-Hermitian (NH) band structures, photonic systems, and classical driven-dissipative systems.

Spectral winding ("point-gap") invariants: Given a NH periodic Bloch Hamiltonian $\gamma\subset\mathbb{C}$9 (possibly matrix-valued), the spectral winding number with respect to a reference energy $z_0\notin\gamma$0 is

$z_0\notin\gamma$1

which counts the algebraic number of times the complex spectrum winds around $z_0\notin\gamma$2 in the plane (Li et al., 2020, Ghatak et al., 2019). This invariant is always integral, jumps when the spectrum crosses $z_0\notin\gamma$3 (i.e., at exceptional points), and underlies phase quantization in non-Hermitian topological matter.

Non-Hermitian Floquet (Quantum Walk) Winding: In multi-step non-unitary quantum walks, a winding number is constructed from the real parts of the Bloch vector extracted from the Floquet operator in momentum space, and large winding values (e.g., $z_0\notin\gamma$4) have been observed experimentally via loss measurements in photonic quantum walks (Xiao et al., 2018). The invariants remain robust under partial measurements and enable detection of phase transitions and pseudo-unitary breaking.

Dynamic and Quench-Based Winding: Topological invariants can be extracted from the long-time average ("spin textures") of local observables following a quantum quench. For general two-band models,

$z_0\notin\gamma$5

where the overline denotes long-time averaging (Zhu et al., 2019, Lin et al., 2024). In NH systems, half-integer windings and abrupt phase jumps emerge near exceptional points or in point-gap braided phases, closely tied to braiding degrees of the complex spectrum. These invariants are directly measurable via dynamical post-quench protocols and have been validated in cold atom and circuit QED platforms.

4. Real-space and Amorphous-System Generalizations

Translation symmetry is not required for the existence or quantization of winding invariants. The spectral localizer framework associates to a Hamiltonian $z_0\notin\gamma$6 and position operator $z_0\notin\gamma$7 a "localizer" $z_0\notin\gamma$8, with chiral symmetry $z_0\notin\gamma$9. The small-$$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$0 expansion yields, at leading order, a real-space winding marker

$$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$1

where $$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$2 is the spectral flattening of $$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$3 (Jezequel et al., 31 Jul 2025). This formulation remains valid on disordered, quasicrystalline, or amorphous lattices and connects directly to the integer quantization of spectral localizer indices, thereby bypassing the need for momentum-space topological invariants. Such real-space windings have been used to classify topological phases well beyond strictly periodic systems.

5. Combinatorial and Polynomial Winding Invariants

Combinatorial models, especially those involving loop configurations on tori or graphs, admit rich sets of winding-type invariants encoding their dynamical and structural properties.

Hexagonal lattice color dynamics: For the 3-coloring problem on the periodic hexagonal lattice (torus), each configuration yields three "winding vectors" (for different loop species). Conservative Kempe move dynamics preserves explicit polynomial winding invariants: parity of crossings

$$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$4

quadratic and quartic norms, and higher-degree antipodal invariants (Cépas et al., 2019, Cépas et al., 2020). These invariants fully classify the ergodic sectors of color dynamics up to steric (finite-size) obstruction and correspond, under suitable embedding or immersion, to linking numbers (including those with Arf-Kervaire invariant on self-intersecting tori).

Almost embeddings of graphs: For any finite graph mapping "almost embedding" into the plane, one defines winding-number-type invariants for cycles around points, together with higher-order cyclic and triodic Wu numbers, e.g.,

$$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$5

for $$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$6, and relations to triodic combinations. These invariants connect to deleted-product homology and encode planarity obstructions, parity constraints, and higher-multiplicity intersection data (Alkin et al., 2024).

6. Curve Geometry, Anholonomy, and Higher-Dimensional Winding

Winding and linking numbers have geometric incarnations in terms of the intrinsic differential geometry of space curves and fields on manifolds. The torsion and total twist of the Frenet–Serret frame provide an explicit decomposition:

• 2D winding number as torsion flux:

$$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$7

• 3D winding and linking via anholonomy:

$$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$8

where $$w(\gamma,z_0)=\frac{1}{2\pi i}\int_\gamma\frac{dz}{z-z_0},$$9 are the total twist densities (torsion plus twist angular derivative) along space curves in the manifold (Balakrishnan et al., 2023). These formulae generalize the Berry-phase–based topological invariants in quantum systems and appear also in fluid helicity and magnetic topology.

7. Physical and Algorithmic Significance

Complex winding invariants serve as critical tools for:

• Generalized residue theorems and singular integral evaluation (when poles lie on contours) (Hungerbühler et al., 2018).
• Topological classification of non-Hermitian and Floquet phases: quantized amplification/attenuation, robustness under disorder, and direct connection to observable plateaus in measurable quantities (Li et al., 2020, Xiao et al., 2018, Lin et al., 2024).
• Algorithmic root-counting and stability analysis, as in certified computation of complex root numbers and Routh-Hurwitz stability (Li et al., 2018).
• Ergodic sector classification and obstruction in coloring/loop models, fully captured by sets of polynomial winding invariants (Cépas et al., 2019, Cépas et al., 2020).
• Graph planarity and topological obstructions via cycle windings and Wu numbers (Alkin et al., 2024).
• Unified geometric and topological structure in both classical and quantum settings, with geometric phase and anholonomy connections (Balakrishnan et al., 2023).

The broad applicability and flexibility of complex winding invariants, in their various analytic, algebraic, combinatorial, and geometric guises, have made them fundamental in modern mathematical physics and topological classification frameworks across diverse settings.