The simplest and most important case is the degree of a continuous map from the n {\displaystyle n} -sphere S n {\displaystyle S^{n}} to itself (in the case n = 1 {\displaystyle n=1} , this is called the winding number):
Let f : S n → S n {\displaystyle f\colon S^{n}\to S^{n}} be a continuous map. Then f {\displaystyle f} induces a pushforward homomorphism f ∗ : H n ( S n ) → H n ( S n ) {\displaystyle f_{*}\colon H_{n}\left(S^{n}\right)\to H_{n}\left(S^{n}\right)} , where H n ( ⋅ ) {\displaystyle H_{n}\left(\cdot \right)} is the n {\displaystyle n} th homology group. Considering the fact that H n ( S n ) ≅ Z {\displaystyle H_{n}\left(S^{n}\right)\cong \mathbb {Z} } , we see that f ∗ {\displaystyle f_{*}} must be of the form f ∗ : x ↦ α x {\displaystyle f_{*}\colon x\mapsto \alpha x} for some fixed α ∈ Z {\displaystyle \alpha \in \mathbb {Z} } . This α {\displaystyle \alpha } is then called the degree of f {\displaystyle f} .
Let X and Y be closed connected oriented m-dimensional manifolds. Poincare duality implies that the manifold's top homology group is isomorphic to Z. Choosing an orientation means choosing a generator of the top homology group.
A continuous map f : X →Y induces a homomorphism f∗ from Hm(X) to Hm(Y). Let [X], resp. [Y] be the chosen generator of Hm(X), resp. Hm(Y) (or the fundamental class of X, Y). Then the degree of f is defined to be f*([X]). In other words,
If y in Y and f −1(y) is a finite set, the degree of f can be computed by considering the m-th local homology groups of X at each point in f −1(y). Namely, if f − 1 ( y ) = { x 1 , … , x m } {\displaystyle f^{-1}(y)=\{x_{1},\dots ,x_{m}\}} , then
In the language of differential topology, the degree of a smooth map can be defined as follows: If f is a smooth map whose domain is a compact manifold and p is a regular value of f, consider the finite set
By p being a regular value, in a neighborhood of each xi the map f is a local diffeomorphism. Diffeomorphisms can be either orientation preserving or orientation reversing. Let r be the number of points xi at which f is orientation preserving and s be the number at which f is orientation reversing. When the codomain of f is connected, the number r − s is independent of the choice of p (though n is not!) and one defines the degree of f to be r − s. This definition coincides with the algebraic topological definition above.
The same definition works for compact manifolds with boundary but then f should send the boundary of X to the boundary of Y.
One can also define degree modulo 2 (deg2(f)) the same way as before but taking the fundamental class in Z2 homology. In this case deg2(f) is an element of Z2 (the field with two elements), the manifolds need not be orientable and if n is the number of preimages of p as before then deg2(f) is n modulo 2.
Integration of differential forms gives a pairing between (C∞-)singular homology and de Rham cohomology: ⟨ c , ω ⟩ = ∫ c ω {\textstyle \langle c,\omega \rangle =\int _{c}\omega } , where c {\displaystyle c} is a homology class represented by a cycle c {\displaystyle c} and ω {\displaystyle \omega } a closed form representing a de Rham cohomology class. For a smooth map f: X →Y between orientable m-manifolds, one has
where f∗ and f∗ are induced maps on chains and forms respectively. Since f∗[X] = deg f · [Y], we have
for any m-form ω on Y.
If Ω ⊂ R n {\displaystyle \Omega \subset \mathbb {R} ^{n}} is a bounded region, f : Ω ¯ → R n {\displaystyle f:{\bar {\Omega }}\to \mathbb {R} ^{n}} smooth, p {\displaystyle p} a regular value of f {\displaystyle f} and p ∉ f ( ∂ Ω ) {\displaystyle p\notin f(\partial \Omega )} , then the degree deg ( f , Ω , p ) {\displaystyle \deg(f,\Omega ,p)} is defined by the formula
where D f ( y ) {\displaystyle Df(y)} is the Jacobian matrix of f {\displaystyle f} in y {\displaystyle y} .
This definition of the degree may be naturally extended for non-regular values p {\displaystyle p} such that deg ( f , Ω , p ) = deg ( f , Ω , p ′ ) {\displaystyle \deg(f,\Omega ,p)=\deg \left(f,\Omega ,p'\right)} where p ′ {\displaystyle p'} is a point close to p {\displaystyle p} . The topological degree can also be calculated using a surface integral over the boundary of Ω {\displaystyle \Omega } ,3 and if Ω {\displaystyle \Omega } is a connected n-polytope, then the degree can be expressed as a sum of determinants over a certain subdivision of its facets.4
The degree satisfies the following properties:5
These properties characterise the degree uniquely and the degree may be defined by them in an axiomatic way.
In a similar way, we could define the degree of a map between compact oriented manifolds with boundary.
The degree of a map is a homotopy invariant; moreover for continuous maps from the sphere to itself it is a complete homotopy invariant, i.e. two maps f , g : S n → S n {\displaystyle f,g:S^{n}\to S^{n}\,} are homotopic if and only if deg ( f ) = deg ( g ) {\displaystyle \deg(f)=\deg(g)} .
In other words, degree is an isomorphism between [ S n , S n ] = π n S n {\displaystyle \left[S^{n},S^{n}\right]=\pi _{n}S^{n}} and Z {\displaystyle \mathbf {Z} } .
Moreover, the Hopf theorem states that for any n {\displaystyle n} -dimensional closed oriented manifold M, two maps f , g : M → S n {\displaystyle f,g:M\to S^{n}} are homotopic if and only if deg ( f ) = deg ( g ) . {\displaystyle \deg(f)=\deg(g).}
A self-map f : S n → S n {\displaystyle f:S^{n}\to S^{n}} of the n-sphere is extendable to a map F : B n + 1 → S n {\displaystyle F:B_{n+1}\to S^{n}} from the n+1-ball to the n-sphere if and only if deg ( f ) = 0 {\displaystyle \deg(f)=0} . (Here the function F extends f in the sense that f is the restriction of F to S n {\displaystyle S^{n}} .)
There is an algorithm for calculating the topological degree deg(f, B, 0) of a continuous function f from an n-dimensional box B (a product of n intervals) to R n {\displaystyle \mathbb {R} ^{n}} , where f is given in the form of arithmetical expressions.6 An implementation of the algorithm is available in TopDeg - a software tool for computing the degree (LGPL-3).
Brouwer, L. E. J. (1911). "Über Abbildung von Mannigfaltigkeiten". Mathematische Annalen. 71 (1): 97–115. doi:10.1007/bf01456931. S2CID 177796823. /wiki/Luitzen_Egbertus_Jan_Brouwer ↩
Siegberg, Hans Willi (1981). "Some Historical Remarks Concerning Degree Theory". The American Mathematical Monthly. 88 (2): 125. doi:10.2307/2321135. /wiki/Doi_(identifier) ↩
Polymilis, C.; Servizi, G.; Turchetti, G.; Skokos, Ch.; Vrahatis, M. N. (May 2003). "Locating Periodic Orbits by Topological Degree Theory". Libration Point Orbits and Applications: 665–676. arXiv:nlin/0211044. doi:10.1142/9789812704849_0031. ISBN 978-981-238-363-1. 978-981-238-363-1 ↩
Stynes, Martin (June 1979). "A simplification of Stenger's topological degree formula" (PDF). Numerische Mathematik. 33 (2): 147–155. doi:10.1007/BF01399550. Retrieved 21 September 2024. https://cs.nyu.edu/~exact/pap/mesh/collection/stynes_simplificationTopDeg79.pdf ↩
Dancer, E. N. (2000). Calculus of Variations and Partial Differential Equations. Springer-Verlag. pp. 185–225. ISBN 3-540-64803-8. 3-540-64803-8 ↩
Franek, Peter; Ratschan, Stefan (2015). "Effective topological degree computation based on interval arithmetic". Mathematics of Computation. 84 (293): 1265–1290. arXiv:1207.6331. doi:10.1090/S0025-5718-2014-02877-9. ISSN 0025-5718. S2CID 17291092. /wiki/ArXiv_(identifier) ↩