ScalingStacks

8.2. Monotone regularization of θ -psh functions [01HG]

Original official author HTML, exact retained edition. Historical TeX conversion verdicts remain unchanged. Cited-edition alignment and mathematical self-containment are not assessed.

Complete original source context · Original author HTML

8.2. Monotone regularization of θ\theta-psh functions

By our definition, the set 𝒟⁡(X)∩PSH⁡(X,θ)\mathcal{D}(X)\cap\PSH(X,\theta) of θ\theta-psh model functions is dense in PSH⁡(X,θ)\PSH(X,\theta) with respect to its topology of uniform convergence on dual complexes. This property may be seen as an analogue of the fact that every θ\theta-psh function is a L1L^{1}-limit of smooth θ\theta-psh functions in the complex case, which follows from the much more useful fact that every θ\theta-psh function is a decreasing limit of smooth θ\theta-psh functions [Dem92]. The next result gives an analogue of this monotone regularization theorem in our context.

Theorem 8.7.

For each θ\theta-psh function φ\varphi, there exists a decreasing net (φi)i∈I(\varphi_{i})_{i\in I} of θ\theta-psh model functions that converges pointwise on XX to φ\varphi.

One may hope that there is in fact a decreasing sequence (φm)m=1∞(\varphi_{m})_{m=1}^{\infty} of θ\theta-psh model functions converging to φ\varphi. We will prove that this is the case in the companion paper [BFJ], using Theorem 8.7 and capacity estimates.

As a consequence of Theorem 8.7, we get at any rate the following version of the Demailly-Richberg regularization theorem.

Corollary 8.8.

Every continuous θ\theta-psh function φ\varphi is the uniform limit on XX of a sequence (φm)m∈𝐍(\varphi_{m})_{m\in\mathbf{N}} of θ\theta-psh model functions.

Proof.

By Theorem 8.7 there exists a decreasing net (φj)(\varphi_{j}) of θ\theta-psh model functions converging pointwise to φ\varphi. For each ε>0\varepsilon>0 the compact set XX is the increasing union of the open sets {φj<φ+ε}\{\varphi_{j}<\varphi+\varepsilon\}, hence φj<φ+ε\varphi_{j}<\varphi+\varepsilon for some jj (Dini’s lemma). It follows that φ\varphi lies in the closure of 𝒟⁡(X)∩PSH⁡(X,θ)\mathcal{D}(X)\cap\PSH(X,\theta) in C0​(X)C^{0}(X) with respect to the topology of uniform convergence. Since the latter is defined by a norm, the result follows. ∎

The proof of Theorem 8.7 reduces immediately to Theorem 8.3, in view of the following elementary result.

Lemma 8.9.

The following properties are equivalent.

  • (i)

    Every θ\theta-psh function φ\varphi is the pointwise limit of a decreasing net of θ\theta-psh model functions.

  • (ii)

    For each u∈C0​(X)u\in C^{0}(X) we have

    Pθ(u)=sup{φ∣φ∈𝒟(X)∩PSH(X,ω),φ≤u on X}.P_{\theta}(u)=\sup\left\{\varphi\mid\varphi\in\mathcal{D}(X)\cap\PSH(X,\omega),\,\varphi\leq u\text{ on }X\right\}.
  • (iii)

    For each u∈C0​(X)u\in C^{0}(X) Pθ​(u)P_{\theta}(u) is a uniform limit of θ\theta-psh model functions.

Proof.

(i)⟹\Longrightarrow(ii). Let u∈C0​(X)u\in C^{0}(X). By (i) there exists a decreasing net (φj)(\varphi_{j}) of θ\theta-psh model functions converging pointwise to Pθ​(u)P_{\theta}(u). Since Pθ​(u)≤uP_{\theta}(u)\leq u, we see that the compact set XX is for each ε>0\varepsilon>0 the increasing union of the open sets {φj<u+ε}\{\varphi_{j}<u+\varepsilon\}, hence φj<u+ε\varphi_{j}<u+\varepsilon for some jj. Since φj−ε\varphi_{j}-\varepsilon is θ\theta-psh and dominated by uu, we get φj−ε≤Pθ​(u)\varphi_{j}-\varepsilon\leq P_{\theta}(u) by definition of the envelope, which proves (ii).

(ii)⟹\Longrightarrow(iii). Since the set of φ∈𝒟⁡(X)∩PSH⁡(X,ω)\varphi\in\mathcal{D}(X)\cap\PSH(X,\omega) such that φ≤u\varphi\leq u is stable by max, (ii) shows that we can construct an increasing family φj∈𝒟⁡(X)∩PSH⁡(X,ω)\varphi_{j}\in\mathcal{D}(X)\cap\PSH(X,\omega) converging pointwise to Pθ​(u)P_{\theta}(u). But Pθ​(u)−φjP_{\theta}(u)-\varphi_{j} is usc for each jj, and Dini’s lemma therefore shows that the convergence is uniform on XX.

(iii)⟹\Longrightarrow(i). Let φ\varphi be a θ\theta-psh function. We first claim that for each x∈Xx\in X we have

φ(x)=inf{ψ(x)∣ψ∈𝒟(X)∩PSH(X,ω),ψ≥φ}.\varphi(x)=\inf\left\{\psi(x)\mid\psi\in\mathcal{D}(X)\cap\PSH(X,\omega),\psi\geq\varphi\right\}.

Indeed, given ε>0\varepsilon>0 there exists u∈C0​(X)u\in C^{0}(X) such that u≥φu\geq\varphi and u⁡(x)≤φ⁡(x)+εu(x)\leq\varphi(x)+\varepsilon, simply because φ\varphi is usc. Since φ\varphi is θ\theta-psh we have φ≤Pθ​(u)\varphi\leq P_{\theta}(u). By (iii) may then find ψ∈𝒟⁡(X)∩PSH⁡(X,ω)\psi\in\mathcal{D}(X)\cap\PSH(X,\omega) such that Pθ​(u)≤ψ≤Pθ​(u)+εP_{\theta}(u)\leq\psi\leq P_{\theta}(u)+\varepsilon. We thus have ψ≥φ\psi\geq\varphi and ψ⁡(x)≤φ⁡(x)+2​ε\psi(x)\leq\varphi(x)+2\varepsilon, and the claim follows.

Now consider the set II of all ψ∈𝒟⁡(X)∩PSH⁡(X,ω)\psi\in\mathcal{D}(X)\cap\PSH(X,\omega) such that ψ>φ\psi>\varphi on XX. Note that this condition implies that ψ≥φ+ε\psi\geq\varphi+\varepsilon for some ε>0\varepsilon>0, since φ−ψ\varphi-\psi is usc. We claim that II is a directed set, which will conclude the proof. To see this, let ψ1,ψ2∈I\psi_{1},\psi_{2}\in I and choose ε>0\varepsilon>0 such that min⁡{ψ1,ψ2}≥φ+3​ε\min\{\psi_{1},\psi_{2}\}\geq\varphi+3\varepsilon. We then also have P⁡(min⁡{ψ1,ψ2})≥φ+3​εP(\min\{\psi_{1},\psi_{2}\})\geq\varphi+3\varepsilon. By (iii) we find ψ3∈𝒟⁡(X)∩PSH⁡(X,ω)\psi_{3}\in\mathcal{D}(X)\cap\PSH(X,\omega) such that |ψ3−P⁡(min⁡{ψ1,ψ2})|≤ε|\psi_{3}-P(\min\{\psi_{1},\psi_{2}\})|\leq\varepsilon. Then φ+ε≤ψ3−ε≤min⁡{ψ1,ψ2}\varphi+\varepsilon\leq\psi_{3}-\varepsilon\leq\min\{\psi_{1},\psi_{2}\}, which concludes the proof. ∎

Original mathematics by the credited authors. Source-backed reader collection; mathematical self-containment is not assessed.