Definition The upper-end edge quotient rule
✓ definition formalized

The upper-end edge quotient rule

For a labelled edge \(e:x\to y\), with label \(\alpha_e\in A\), the upper-quotient condition says that \[ \rho^+_e:B^y\longrightarrow B^e \] is surjective and \[ \ker\rho^+_e=\alpha_e B^y. \] The induced isomorphism \(B^y/\alpha_eB^y\simeq B^e\) sends the class of \(v\) to \(\rho^+_e(v)\). This identifies the edge module together with its map from the upper stalk.

The label annihilates the edge module. Indeed, write an edge vector as \(\rho^+_e(v)\); then \[ \alpha_e\rho^+_e(v)=\rho^+_e(\alpha_ev)=0. \] Thus \(B^e\) is naturally a module over \(A/(\alpha_e)\).

If \(B^y\) has a finite free basis over \(A\), its basis classes give a basis of \(B^e\) over \(A/(\alpha_e)\).

The degree-one exact sequence

Suppose \(\alpha_e\) is homogeneous of degree one and multiplication by it on \(B^y\) is injective. The shift convention is \[ M(d)_n=M_{n-d}. \] Hence multiplication gives a degree-zero map \(B^y(1)\to B^y\): a degree-\(n\) source vector is originally in \(B^y_{n-1}\), and its product with \(\alpha_e\) lies in \(B^y_n\).

The exact sequence is therefore \[ 0\longrightarrow B^y(1) \xrightarrow{\ \alpha_e\ }B^y \xrightarrow{\ \rho^+_e\ }B^e \longrightarrow0. \] Its image and kernel agree by the quotient condition. Suppose the graded pieces are finite-dimensional over \(k\). Write \[ H_M(q)=\sum_{n\ge0}\dim_k(M_n)q^n \] for the Hilbert series, or its value at a positive \(q\) where the sum converges. In each degree, exactness makes dimensions add, so \[ H_{B^e}(q)=H_{B^y}(q)-qH_{B^y}(q) =(1-q)H_{B^y}(q). \]

For \(A=k[\alpha]\) and \(B^y=A\), the quotient is \(k\), concentrated in degree zero. The identity reads \(1=(1-q)/(1-q)\).

The lower restriction \(\rho^-_e\) describes how values at \(x\) enter this quotient. Its restrictions to outgoing kernels give the lower images used in the comparison.

References

The local definitions and constructions are listed below.

Lean formalization

OAI.KLInvariance.MomentGraph.GradedSheaf.UpperQuotient: GradedSheaves.lean:401.

OAI.KLInvariance.MomentGraph.Sheaf.UpperQuotient: GradedSheaves.lean:963.

OAI.KLInvariance.MomentGraph.GradedSheaf.UpperQuotient.annihilates: Support.lean:871.

Used by