Bourbaki

Proposition 8

Tag 0CSV, permanent. Cite it as [Bourbaki, Tag 0CSV].

Commutative Algebra, chapter VIII, § 7. Multiplicités

Let δ>0\delta > 0 be an integer, x an element of qδq^\delta, ξ\xi its class in grδ(A)=qδ/qδ+1\mathrm{gr}_\delta(A) = q^\delta/q^{\delta+1} and φ\varphi the multiplication by ξ\xi in the gr(A)-module gr(M).

a) The dimension of the A-module M/xMM/xM is equal to dim⁡A(M)\dim_A(M) or to dim⁡A(M)−1\dim_A(M) - 1. In the second case, one has eq(M/xM)≥δeq(M)e_q(M/xM) \geq \delta e_q(M).

b) Suppose that one has dim⁡A(M)≥1\dim_A(M) \geq 1 and that the kernel of φ\varphi is of finite length over A/qA/q. Then one has dim⁡A(M/xM)=dim⁡A(M)−1\dim_A(M/xM) = \dim_A(M) - 1. Moreover: (i) If dim⁡A(M)>1\dim_A(M) > 1, one has eq(M/xM)=δeq(M)e_q(M/xM) = \delta e_q(M). (ii) If dim⁡A(M)=1\dim_A(M) = 1, one has for every integer n≥0n \geq 0

nδeq(M)≤longA(M/xnM)≤nδeq(M)+longA/q(Ker φn), n \delta e_q(M) \leq \mathrm{long}_A(M/x^nM) \leq n \delta e_q(M) + \mathrm{long}_{A/q}(\mathrm{Ker}\ \varphi^n),
where φn\varphi^n is the n-th iterate of the endomorphism φ\varphi, and
δeq(M)=exA(M)≤longA(M/xM). \delta e_q(M) = e_{xA}(M) \leq \mathrm{long}_A(M/xM).

Put M′′=M/xMM'' = M/xM; consider the Hilbert-Samuel series HM=HM,qH_M = H_{M,q} and HM′′=HM′′,qH_{M''} = H_{M'',q}, together with the Poincaré series P(T)=∑n≥0longA/q(Ker φn)⋅TnP(T) = \sum_{n \geq 0} \mathrm{long}_{A/q}(\mathrm{Ker}\ \varphi_n) \cdot T^n. By § 4, No. 3, th. 2 and No. 4, remark 2, one has

HM(T)=(1−T)−dRM(T),HM′′(T)=(1−T)−d′′RM′′(T), H_M(T) = (1 - T)^{-d} R_M(T), \quad H_{M''}(T) = (1 - T)^{-d''} R_{M''}(T),

with d=dim⁡A(M)d = \dim_A(M), d′′=dim⁡A(M′′)d'' = \dim_A(M''), RMR_M and RM′′R_{M''} in Z[T]\mathbf{Z}[T], RM(1)=eq(M)R_M(1) = e_q(M), RM′′(1)=eq(M′′)R_{M''}(1) = e_q(M''). By lemma 6 of § 4, No. 5, one has in Z((T))\mathbf{Z}((T)) the inequalities

(1−Tδ)HM(1)≤HM′′(1)≤(1−Tδ)HM(1)+TδP(1). (1 - T^\delta) H_M^{(1)} \leq H_{M''}^{(1)} \leq (1 - T^\delta) H_M^{(1)} + T^\delta P^{(1)} .

Putting R(T)=(1−Tδ)/(1−T)=1+T+⋯+Tδ−1R(T) = (1 - T^\delta)/(1 - T) = 1 + T + \cdots + T^{\delta-1}, this may also be written

(1)

(1−T)−dR(T)RM(T)≤(1−T)−d′′−1RM′′(T)≤ (1 - T)^{-d} R(T) R_M(T) \leq (1 - T)^{-d''-1} R_{M''}(T) \leq
≤(1−T)−dR(T)RM(T)+(1−T)−1TδP(T). \leq (1 - T)^{-d} R(T) R_M(T) + (1 - T)^{-1} T^\delta P(T) .

By Lemma 2 of § 4, No. 1, the first inequality (1) implies either d′′≥dd'' \geq d, or d′′=d−1d'' = d - 1 and R(1)RM(1)≤RM′′(1)R(1) R_M(1) \leq R_{M''}(1), that is to say δeq(M)≤eq(M′′)\delta e_q(M) \leq e_q(M''). This proves a), since d′′≤dd'' \leq d.

Under the assumption of b), one has P(T)∈Z[T]P(T) \in \mathbf{Z}[T] and P(1)=long⁡A(Ker⁡φ)P(1) = \operatorname{long}_A(\operatorname{Ker} \varphi). The second inequality (1) is written

(1−T)−d′′−1RM′′(T)≤(1−T)−d(R(T)RM(T)+Tδ(1−T)d−1P(T)). (1 - T)^{-d''-1} R_{M''}(T) \leq (1 - T)^{-d}(R(T) R_M(T) + T^\delta (1 - T)^{d-1} P(T)) .

Suppose that one has d>1d > 1; then Lemma 2 of § 4, No. 1 yields d′′+1≤dd'' + 1 \leq d, whence d′′=d−1d'' = d - 1 by part a) of the proof; one then has

RM′′(1)≤R(1).RM(1) R_{M''}(1) \leq R(1). R_M(1)

(loc. cit.), whence (i).

Now suppose d=1d = 1. By loc. cit., one has d′′=0d'' = 0 and

RM′′(1)≤R(1).RM(1)+P(1). R_{M''}(1) \leq R(1). R_M(1) + P(1) .

Consequently, M′′M'' is of finite length equal to eq(M′′)=RM′′(1)e_q(M'') = R_{M''}(1), and one obtains

(2)

δeq(M)≤long⁡A(M/xM)≤δeq(M)+long⁡A(Ker⁡φ). \delta e_q(M) \leq \operatorname{long}_A(M/xM) \leq \delta e_q(M) + \operatorname{long}_A(\operatorname{Ker} \varphi) .

Let n≥1n \geq 1 be an integer. Replace xx by xnx^n in (2); one therefore has

(3)

nδeq(M)≤long⁡A(M/xnM)≤nδeq(M)+long⁡A(Ker⁡φn). n \delta e_q(M) \leq \operatorname{long}_A(M/x^nM) \leq n \delta e_q(M) + \operatorname{long}_A(\operatorname{Ker} \varphi^n) .

It is immediate that the submodules Ker⁡φn\operatorname{Ker} \varphi^n of the noetherian gr(A)-module gr(M) form an increasing sequence and hence are stationary, and that each of them is of finite length over A/qA/q. Dividing by n≥1n \geq 1 in inequality (3) and making nn tend to +∞+ \infty, one finds exA(M)=δeq(M)e_{xA}(M) = \delta e_q(M) by definition of exA(M)e_{xA}(M).

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Nothing.

Cites

Nothing.

Translated with machine help, not yet reviewed by a person.