ScalingStacks

7.4.5. Generation

We equip ZZ with a metric. Given ξ\xi a path in ZZ, we denote by |ξ||\xi| its length. Given ζ\zeta a homotopy class of paths in ZZ, we put |ζ|=|ξ||\zeta|=|\xi|, where ξ\xi is a minimal path in ζ\zeta. Given θ:I→J\theta:I\to J a braid in ZZ, we put |θ|=∑s∈I|θs||\theta|=\sum_{s\in I}|\theta_{s}|.

Let MM be a finite subset of ZZ.

0PBJ

Lemma 7.4.25. Let θ∈Hom𝒮M∙​(Z)⁡(I,J)\theta\in\operatorname{Hom}\nolimits_{{\mathcal{S}}^{\bullet}_{M}(Z)}(I,J) and θ′∈Hom𝒮M∙​(Z)⁡(I′,I)\theta^{\prime}\in\operatorname{Hom}\nolimits_{{\mathcal{S}}^{\bullet}_{M}(Z)}(I^{\prime},I) such that θ∘θ′\theta\circ\theta^{\prime} is a braid. Let I0I_{0} be a finite subset of M∖(I∪I′∪J)M\setminus(I\cup I^{\prime}\cup J).

If |θ|+|θ′|=|θ∘θ′||\theta|+|\theta^{\prime}|=|\theta\circ\theta^{\prime}|, then (θ⊠idI0)⋅(θ′⊠idI0)=(θ⋅θ′)⊠idI0(\theta\boxtimes\operatorname{id}\nolimits_{I_{0}})\cdot(\theta^{\prime}\boxtimes\operatorname{id}\nolimits_{I_{0}})=(\theta\cdot\theta^{\prime})\boxtimes\operatorname{id}\nolimits_{I_{0}}.

0PBK

Proof. Let s′∈I′s^{\prime}\in I^{\prime}. Since |θθ′​(s′)|+|θs′′|=|θθ′​(s′)∘θs′′||\theta_{\theta^{\prime}(s^{\prime})}|+|\theta^{\prime}_{s^{\prime}}|=|\theta_{\theta^{\prime}(s^{\prime})}\circ\theta^{\prime}_{s^{\prime}}|, it follows that i⁡(θθ′​(s′),idi)+i⁡(θs′′,idi)=i⁡(θθ′​(s′)∘θs′′,idi)i(\theta_{\theta^{\prime}(s^{\prime})},\operatorname{id}\nolimits_{i})+i(\theta^{\prime}_{s^{\prime}},\operatorname{id}\nolimits_{i})=i(\theta_{\theta^{\prime}(s^{\prime})}\circ\theta^{\prime}_{s^{\prime}},\operatorname{id}\nolimits_{i}) for all i∈I0i\in I_{0}. As a consequence,

i⁡((θ∘θ′)⊠idI0)−i⁡(θ⊠idI0)−i⁡(θ′⊠idI0)=i⁡(θ∘θ′)−i⁡(θ)−i⁡(θ′).i((\theta\circ\theta^{\prime})\boxtimes\operatorname{id}\nolimits_{I_{0}})-i(\theta\boxtimes\operatorname{id}\nolimits_{I_{0}})-i(\theta^{\prime}\boxtimes\operatorname{id}\nolimits_{I_{0}})=i(\theta\circ\theta^{\prime})-i(\theta)-i(\theta^{\prime}).

The lemma follows now from Lemma 7.4.9. ∎

Let θ∈Hom𝒮M∙​(Z)⁡(I,J)\theta\in\operatorname{Hom}\nolimits_{{\mathcal{S}}^{\bullet}_{M}(Z)}(I,J) be a non-zero braid.

Let I0={i∈I|θi=idi}I_{0}=\{i\in I\ |\ \theta_{i}=\operatorname{id}\nolimits_{i}\}. Let i∈I∖I0i\in I\setminus I_{0}. There is a (unique) decomposition θi=βi⋅αi\theta_{i}=\beta^{i}\cdot\alpha^{i} in 𝒮∙​(Z,1){\mathcal{S}}^{\bullet}(Z,1) with

  • •

    αi​(1)∈M∖I0\alpha^{i}(1)\in M\setminus I_{0}

  • •

    |θi|=|αi|+|βi||\theta_{i}|=|\alpha^{i}|+|\beta^{i}|

  • •

    given a minimal path ξ\xi in αi\alpha^{i}, we have ξ⁡((0,1))∩M⊂I0\xi((0,1))\cap M\subset I_{0}.

We define a quiver Γ⁡(θ)\Gamma(\theta) with vertex set I∖I0I\setminus I_{0}. There is an arrow i→i′i\to i^{\prime} if αi​(1)=i′\alpha^{i}(1)=i^{\prime}.

Note that there is at most one arrow with a given source (that arrow can be a loop).

0PBL

Lemma 7.4.26. Let I′I^{\prime} be a non-empty finite subset of I∖I0I\setminus I_{0} such that

  • •

    if there is an arrow i→i′i\to i^{\prime} in Γ⁡(θ)\Gamma(\theta) with i∈I′i\in I^{\prime}, then i′∈I′i^{\prime}\in I^{\prime}

  • •

    given i≠i′∈I′i\neq i^{\prime}\in I^{\prime}, we have αi​(1)≠αi′​(1)\alpha^{i}(1)\neq\alpha^{i^{\prime}}(1).

There is a (unique) decomposition θ=θu⋅u\theta=\theta^{u}\cdot u in 𝒮M∙​(Z){\mathcal{S}}_{M}^{\bullet}(Z), where |θ|=|θu|+|u||\theta|=|\theta^{u}|+|u| and

ui={αi if ​i∈I′idi otherwise.u_{i}=\begin{cases}\alpha^{i}&\text{ if }i\in I^{\prime}\\ \operatorname{id}\nolimits_{i}&\text{ otherwise.}\end{cases}
0PBM

Proof. Note that the second assumption on I′I^{\prime} shows that the full subquiver of Γ⁡(θ)\Gamma(\theta) with vertex set I′I^{\prime} is a disjoint union of oriented lines and oriented circles.

Let i1≠i2∈Ii_{1}\neq i_{2}\in I. If i1,i2∈I∖I′i_{1},i_{2}\in I\setminus I^{\prime}, then u⁡(i1)≠u⁡(i2)u(i_{1})\neq u(i_{2}). Assume now i1∈I′i_{1}\in I^{\prime} and i2∈I∖I′i_{2}\in I\setminus I^{\prime}. Since i1→i2i_{1}\to i_{2} is not an arrow of the quiver, we have u⁡(i1)≠i2u(i_{1})\neq i_{2}, hence u⁡(i1)≠u⁡(i2)u(i_{1})\neq u(i_{2}). Finally if i1,i2∈I′i_{1},i_{2}\in I^{\prime}, then u⁡(i1)≠u⁡(i2)u(i_{1})\neq u(i_{2}). We have shown that uu is a braid.

Note that there is a (unique) decomposition θ=θu∘u\theta=\theta^{u}\circ u with |θ|=|θu|+|u||\theta|=|\theta^{u}|+|u|. In order to show that θu⋅u≠0\theta^{u}\cdot u\neq 0, we can replace θ\theta by θ|I∖I0\theta_{|I\setminus I_{0}} and MM by M∖I0M\setminus I_{0}, thanks to Lemma 7.4.25. So, we assume now that I0=∅I_{0}=\emptyset.

Let q:Z~→Zq:\tilde{Z}\to Z be a non-singular cover of ZZ. Let M~=q−1​(M)\tilde{M}=q^{-1}(M). Let θ~:I~→J~\tilde{\theta}:\tilde{I}\to\tilde{J} be the unique lift of θ\theta to Z~\tilde{Z}. We have a decomposition θ~i=β~i⋅α~i\tilde{\theta}_{i}=\tilde{\beta}^{i}\cdot\tilde{\alpha}^{i} for i∈I~i\in\tilde{I} and q⁡(α~i)=αq⁡(i)q(\tilde{\alpha}^{i})=\alpha^{q(i)}.

Let I~′=q−1​(I′)∩I~\tilde{I}^{\prime}=q^{-1}(I^{\prime})\cap\tilde{I}. Note that qq induces a morphism of quivers Γ⁡(θ~)→Γ⁡(θ)\Gamma(\tilde{\theta})\to\Gamma(\theta), hence I~′\tilde{I}^{\prime} satisfies the assumptions of the lemma and we have a decomposition θ~=θ~u~∘u~\tilde{\theta}=\tilde{\theta}^{\tilde{u}}\circ\tilde{u}. Since q⁡(u~)=uq(\tilde{u})=u, it follows that if the lemma holds for θ~\tilde{\theta}, then it holds for θ\theta.

We assume now that ZZ is non-singular. If the lemma holds for connected components of ZZ, it will hold for ZZ, hence it is enough to prove the lemma for ZZ connected. Assume now ZZ is connected. There is an injective morphism of curves f:Z→S1f:Z\to S^{1}, where S1S^{1} is unoriented. It the lemma holds for S1S^{1}, it holds for ZZ.

We assume finally that Z=S1Z=S^{1} unoriented. Let i1≠i2∈I′i_{1}\neq i_{2}\in I^{\prime} such that i⁡(ui1,ui2)≠0i(u_{i_{1}},u_{i_{2}})\neq 0. Note that ui1u_{i_{1}} and ui2u_{i_{2}} have opposite directions and i⁡(ui1,ui2)=1i(u_{i_{1}},u_{i_{2}})=1. Furthermore, θir\theta_{i_{r}} has the same direction as uiru_{i_{r}}, hence i⁡(θi1,θi2)=i⁡((θu)i1,(θu)i2)+1i(\theta_{i_{1}},\theta_{i_{2}})=i((\theta^{u})_{i_{1}},(\theta^{u})_{i_{2}})+1. Given i1≠i2∈Ii_{1}\neq i_{2}\in I with i1∉I′i_{1}{\not\in}I^{\prime}, we have i⁡(ui1,ui2)=0i(u_{i_{1}},u_{i_{2}})=0. It follows from Remark 7.4.11 that θu⋅u≠0\theta^{u}\cdot u\neq 0. This completes the proof of the lemma. ∎

Note that the length of a map in 𝒮M∙​(Z){\mathcal{S}}^{\bullet}_{M}(Z) takes value in a finitely generated submonoid of 𝐑≥0{\mathbf{R}}_{\geq 0}. So, a repeated application of the previous lemma provides a decomposition of any map θ\theta of 𝒮M∙​(Z){\mathcal{S}}_{M}^{\bullet}(Z) as a product θ=un⋯u1\theta=u_{n}\cdots u_{1}, where uiu_{i} is a map uu as in the lemma.

Original mathematics by the credited authors. Source collection and HTML conversion remain in progress.

Andrew Manion, Raphael Rouquier

Original source: arXiv:2009.09627v2