Mayer-Vietoris Theorem for Singular Homology

Subjects: Algebraic Topology
Links: Singular Homology, The Seifert-Van Kampen Theorem, The Zigzag Lemma, Simplicial Complexes, Abstract Simplicial Complexes

Def: Suppose U is an open cover of X. A singular chain c is said to be U-small if every singular simplex that appears in c has image lying entirely in one of the open subsets in U. Let CpU(X) denote the subgroup of Cp(X) consisting of U-small chains, and let HpU(X) denote the homology of the complex CU(X).

Def: If α=A(v0,,vp) is an affine singular p-simplex in some convex set KRm and w is any point in K, we define an affine singular (p+1)-simplex wα called the cone on α from w by $$w* \alpha := w* A(v_0,\dots, v_p) = A(w, v_0,\dots, v_p). $$In other words, wα:Δp+1K is the unique affine simplex that sends e0 to w and whose 0th dace map is equal α. We extend this operator to affine chains by linearity $$w* \left(\sum_i n_i\alpha_i\right) := \sum_i n_i (w *\alpha_i).$$It is not defined for arbitrary singular chains.

Lemma: If c is an affine chain. then $$\partial(wc) + w\partial c= c,$$or equivalently, $$\partial(wc) = c-w\partial c.$$
Def: For any k-simplex σ=[v0,,vk]Rn, we define the barycentre of σ to be the points bσInt σ whose barycentric coordinates are all equal: $$b_\sigma := \sum_{i = 0}^k \frac1{k+1} v_i. $$
Now we define an operator s taking affine p-chains to affine p-chains, called the singular subdivision operator. For p=0, simply set s=id. For p>0, assuming that s has been defined for chains of dimensions less than p, for any affine p-simplex α:ΔpRn we set $$s\alpha := \alpha(b_p)* s (\partial \alpha), $$where bp is the barycentre of Δp, and extend linearly to affine chains.

Lemma: Suppose α:ΔpRn is an affine simplex that is a homeomorphism onto a p-simplex σRn. Let β:ΔpRn be any one of the affine singular p-simplices that appear in the chains sα.

Now we need to extend the singular subdivision operator to arbitrary, not necessarily affine, singular chains. For a singular p-simples σ in any space X, note that σ=σ#ip, where ip:ΔpΔp is the the identity map considered as an affine singular p-simplex in Δp, and σ#:Cp(Δp)Cp(X) is the chain map obtained from the continuous map σ:ΔpX. we define $$s\sigma := \sigma_# (si_p),$$and extend by linearity to all of Cp(X). We can iterate s to obtain operators s2:=ss and more generally sk+1:=sks.

Lemma: The singular subdivision operators s:Cp(X)Cp(X) have the following properties.

The proof of the third fact, actually needs the Lebesgue Number Lemma

Prop: Suppose U is any open cover of X. The inclusion map CU(X)C(X) induces a homology isomorphism HpU(X)Hp(X) for all p.

We are given a space X and two open subsets U,VX whose union is X. There are four inclusion maps

\usepackage{tikz-cd}
\usepackage{amsfonts, amsmath, amssymb}

\begin{document}
\begin{tikzcd}
& U\arrow[dr, "k"] & \\
U \cap V\arrow[ur, "i"]\arrow[dr, "j"] && X \\
&V\arrow[ur,"\ell"']&
\end{tikzcd}
\end{document}

Mayer-Vietoris Theorem: Let X be a topological space, and let U,V be open subsets of whose union is X. Then for each p there is a homomorphism :Hp(X)Hp1(UV) such that the following sequence is exact: $$\cdots \stackrel{\partial_}{\longrightarrow}H_p(U \cap V) \stackrel{i_ \oplus j_}{\longrightarrow} H_p(U) \oplus H_p(V)\stackrel{k_- l_}{\longrightarrow}H_p(X)\stackrel{\partial_}{\longrightarrow} H_{p-1}(U\cap V)\stackrel{i_* \oplus j_*}{\longrightarrow}\cdots$$
Prop: Let X1,,Xk be spaces with nondegenerate base points. For every p>0, then $$H_p(X_1\vee \dots \vee X_k) = H_p(X_1) \oplus \dots \oplus H_p(X_k)$$