• Aucun résultat trouvé

RENORMALIZATION OF GENERALIZED “abc” GASKETS

N/A
N/A
Protected

Academic year: 2022

Partager "RENORMALIZATION OF GENERALIZED “abc” GASKETS"

Copied!
11
0
0

Texte intégral

(1)

ANTONIO-MIHAIL NUIC ˘A

We obtain existence of eigenforms (eigenoperators) for the renormalization map of the so-called generalized “abc” gasket. The method we used is based on Hilbert projective metric approach (developed by Volker Metz). We also com- pute the “highly-symmetric” irreducible eigenoperator of the renormalization map for a particular example, the so-called “affine Sierpinski gasket”.

AMS 2010 Subject Classification: 31XX, 60J45, 28A80.

Key words: post critical finite self-similar structure, Laplacian, Dirichlet form, electrical networks, Hilbert projective metric, renormalization map, eigenform, diffusion process, shortcut procedure.

1. INTRODUCTION

The post critical finite self-similar structures (introduced by Jun Kigami in [4]) are finitely ramified structures. Theaffine nested fractals are connected post critical finite self-similar structures [3].

The renormalization map of a post critical finite self-similar structure is defined on the cone of all possible energy forms (or associated operators) of the structure initial boundary. The existence of a Dirichlet form on a finitely ramified fractal is equivalent to the existence of aneigenform of the renormal- ization map of the structure (Kusuoka – [7]). This eigenvalue problem was studied by T. Lindstrom in [8] for nested fractals and for affine nested fractals in [3]. The existence of the eigenforms of the renormalization map for post critical finite self-similar structures is an open problem; significant results were obtained by V. Metz in [9–11]. If we start with anirreducible eigenformfor the renormalization map of a connected post critical finite self-similar structureF, then we can construct a local regular Dirichlet form onL2(F, µ), where µis a well chosen probability measure on F ([4, 6]).

In the next section, we overview the basic facts on self-similar structures, energy forms and Laplacians on finite sets, Hilbert projective metric on cones;

we remember the most important properties of the renormalization map of a post critical finite self-similar structure (Cf. [9, 11]). In the last section, we present the main result on the existence of eigenforms for a particular class of post critical finite self-similar structures, named generalized “abc” gaskets.

Finally, we compute the irreducible eigenoperator for a particular affine nested fractal,the affine Sierpinski gasket.

MATH. REPORTS15(65),4(2013), 477–487

(2)

2. THE RENORMALIZATION PROBLEM

2.1. SELF-SIMILAR STRUCTURES n

F,{χi}i=1,No

is called self-similar structure if and only if (F, d) is a compact metric space, χi : F −→ F, i = 1, N are injective contractions and F =

N

[

i=1

χi(F) =: Ψ(F). Associated to such a structure is its natu- ral map π : Σ := {1,2, . . . , N}N −→ F. We consider the ramifying set

B := [

1≤i<j≤N

i(F)∩χj(F)), the critical set Γ := π−1(B), the postcritical set P := [

n≥1

σn(Γ) and the initial boundary of the structure V0 := π(P); for w=w1. . . wn. . .∈Σ we consideredσw:=w2. . . wn. . .. The self-similar struc- ture is calledfinitely ramified ifB is finite and post critical finiteifP is finite.

The boundaryV0 generates then-cells: Vn:= [

w∈Wn

Ψn(V0),n≥1.

By restricting the similitudes of an iterated function system on Rd to its attractor (see [2]) we get a self-similar structure, which is called affine nested fractal if it satisfies three axioms: connectivity, symmetry and a so- called nesting axiom (Cf. [3]). An affine nested fractal is a connected post critical finite self-similar structure (Cf. Theorem 5.24 from [1]).

2.2. DIRICHLET FORMS, LAPLACIANS AND CONDUCTANCES Consider a finite setV, the subspace of linear symmetric operatorsLs(V) inRV =n

f

f :V →R o

and the subspace of bilinear symmetric forms Fs(V) in RV. Then Ξ : Ls(V) → Fs(V), Ξ(H) := HH is an isomorphism, where HH(u, v) :=−(u, Hv), u,v∈RV, (u, v) := X

p∈V

u(p)v(p) for u,v∈RV.

Definition 2.1. a) Let H ∈ Fs(V). H can satisfy one of the axioms:

(F.1): H(u, u) ≥ 0, for all u ∈ RV; (F.2): H(u, u) = 0 ⇔ u constant; (F.3):

H(u, u)≤ H(u, u), for allu∈RV (u:= 0∨(1∧u), foru∈RV). Instead of axiom (F.2) we consider (F.2)0: H(1,1) = 0 or (F.2)00: H(u, u) = 0⇒u constant.

b) We consider the cone F120(V) of forms with (F.1), (F.2)0, the cone F1203(V) of forms which satisfy (F.1), (F.2)0 and (F.3) and the cone F123(V) of forms satisfying (F.1), (F.2) and (F.3). The forms in F1203(V) are called Dirichlet forms on V and the forms in F123(V) are calledirreducible Dirichlet

(3)

forms. It can be verified that F1203(V) is polyhedral andF120(V) is not always polyhedral.

c) LetH ∈ Ls(V). H can satisfy one of the properties: (O.1): (Hu, u)≤ 0, for all u ∈RV; (O.2): Hu= 0 ⇔u constant; (O.3): Hpq := (Hχq)(p) ≥0, for all p 6= q ∈ V. Instead of (O.2) we consider: (O.2)0: H1 = 0, that is, X

q∈V

Hpq = 0, for all p ∈ V; (O.2)00: Hu = 0 ⇒ u constant. We denote by L120(V), L1203(V), L12(V), L123(V) the cones of operators H which satisfies (O.1), (O.2)0, or (O.1), (O.2)0 and (O.3), and so on. If H ∈ L12(V), then N := (V, H) is called electrical network on V.

It’s enough for H to satisfy property (O.2)0, in order to have (2.1) HH(h, g) =−(h, Hg) = 1

2 X

p,q∈V

Hpq(h(p)−h(q))(g(p)−g(q)), h, g∈RV. The following Proposition is an extension of Prop. 2.1.3 from [5]:

Proposition 2.2. Ξ(L120(V)) = F120(V); Ξ(L1203(V)) = F1203(V);

Ξ(L12(V)) =F12(V);Ξ(L123(V)) =F1203(V).

Definition 2.3. 1) Aconductanceis a matrixc:= (c(p, q))p,q∈V withcpq :=

c(p, q), having properties (C.1): cpq =cqp, for all p, q∈ V; (C.2): cpp = 0, for allp∈V; (C.3): cpq ≥0, for allp6=q ∈V. We associate the graph Γ := (V, G), G := {{p, q} ⊂ V |cpq >0}. There is an one-to-one correspondence between L1203(V) and conductance matrices.

2) For U $ V and H ∈ Ls(V), consider X : RU −→ RU, Y : RU −→

RV\U, Z : RV\U −→ RV\U, so that H ∼=

X Yt

Y Z

. If H ∈ L12(V), by Lemma 2.1.5 and Theorem 2.1.6 from [5] Z is negative definite, [H]U :=

X −YtZ−1Y is well defined on V and [H]U ∈ L12(U). [H]U is called the trace, orrestriction of H toU. (V, H) and (U,[H]U) are also called equivalent networks. The well known Ohm’s law and ∆−Y transform are particular cases of equivalence.

2.3. HILBERT’S PROJECTIVE METRIC AND THE RENORMALIZATION PROBLEM In this subsection, we fix r:= (r1, . . . , rn), ri >0,S :=n

F,{χi}i=1,No a connected post critical finite self-similar structure, V0 the initial boundary of S, the conesF120(V0) =:F120,F1203(V0) =:F1203,F123(V0) =:F123onV0defined as above and their associated cones of operators.

(4)

If B(V0) :=F1203(V0)−F1203(V0) is the real normed space with respect to the norm ||E||2 := sup

E(u, u)

u∈RV0,||u||V0 = 1 , then it’s easy to verify that

F12\03(V0) =

E ∈F1203

E(χp, χq)<0, p6=q∈V0 ,

F\120(V0) =L120(V0).

Definition 2.4.1. The renormalization mapis defined by:

Λ := Λr:F120(V0)−→F120(V0),Λ := Φ◦Ψ,

where Ψ is the replication “operation” and Φ is the trace “operation”: for A0 ∈ F120(V0), put Ψ(A0) =: A1, with A1(u, u) :=

N

P

i=1

ri−1A0(u◦χi, u◦χi) for all u ∈ RV1; then Ψ(A0) = A1 ∈ F120(V1). For A1 ∈ F120(V1), define Φ(A1) =: A0, where A0(u, u) := inf

A1(v, v)

v ∈RV1, v|V0 =u , u ∈ RV0; thenA0 ∈F120(V0). Proposition 2.2 allows Λ to be defined on operators instead of forms.

2. For K∈ {F120,F1203} and A,B ∈K\{0}, we can define A ∼K B if and only if there existsα,β >0 withαA ≤K B ≤K βA, whereA ≤K Bif and only ifB − A ∈K; then “∼K” is an equivalence relation onK\{0}, the equivalence classes are called parts (subcones of K), K is the most important one and obviously every F1203-part is contained in aF120-part.

IfA ∈F1203, then we define the graph Γ(A) := (V0, G(A)), with G(A) :=

{p, q} ⊂V0

cA(p, q)>0 . If A, B ∈ F1203, from (2.1) written for conduc- tances we get A ∼F

1203 B if and only if Γ(A) = Γ(B). So, we can define Γ(F01203) := Γ(A), forA ∈F01203 andF01203 aF1203 part. Γ(Ψ(F01203)) denotes the graph with vertices V1 associated with Ψ(A), for every A ∈ F01203. Similarly, if A, B ∈ F120\{0}, then A ∼F

120 B if and only if KerA = KerB (KerA is the kernel of the operator associated with A). The following lemma is a very useful tool:

Lemma 2.5 ([12], Prop. 1.15). For A ∈ F1203 and p 6= q ∈ V0 we have cΛ(A)(p, q) > 0 if and only if p and q are Ψ(A)-connected avoiding V0\{p, q}

(cΛ(A) is the conductance matrix associated to Λ(A)).

Definition 2.6.Let K∈ {F120,F1203} andA,B ∈K\{0}.

– for A ∼K B, define m(B|A) := sup α >0

αA ≤ B , M(B|A) := inf

β >0

B ≤βA , h(A,B) := lnM(B|A) m(B|A); – for AK B, consider h(A,B) :=∞,h(0,0) := 0.

h is called the Hilbert projective “metric”.

(5)

In fact, his just a semi-metric onK\{0}. h(A,B) = 0 if and only if there exists α > 0, such that A = αB; h(A,B) = ∞ if and only if A K B; also h(A,B) =h(αA, βB), for allA,B ∈K\{0},α,β >0 (Cf. [9]).

The next proposition summarizes the properties of Λ (Cf. Sections 2 and 4 from [9], Cor. 4, Lemma 7 from [11]). “≤F

120” is denoted simply by “≤”.

Proposition2.7. (a) Λ F120

⊂F120,Λ F1203

⊂F1203,Λ(F123)⊂F1203; Λ is positive homogeneous, ≤-monotone and superadditive;

(b) Λ is non-expansive with respect to h on Λ-invariant F120-parts;

(c) Λ “acts” on F1203-parts and also on F120-parts;

(d)There exists an eigenform in every closed convex Λ-invariant subcone (for example F1203-part) of F1203; if A, B ∈ F120, A ∼F

120 B, Λ(A) = αA, Λ(B) =βB, then α=β.

The renormalization map is studied with respect to subcones which are invariant under specific symmetry groups associated to the structure S:

Definition 2.8.1) Θ is called symmetry group for S if and only if Θ is formed with continuous bijectionsg:F −→F such thatg(V0)⊂V0, for every g∈Θ and if for all i∈ {1, . . . , N} and g∈Θ, there existsj ∈ {1, . . . , N} and g0∈Θ withg◦χij ◦g0.

2) Consider A ∈Fs(V0), H ∈ Ls(V0), c a conductance matrix onV0 and r= (r1, . . . , rN) withri >0,i= 1, N. Then: Ais called Θ-invariantif and only if for allθ∈Θ andu,v∈RV0,A(u◦θ, v◦θ) =A(u, v);His called Θ-invariant if and only if for allθ∈Θ andp,q ∈V0,Hpq =Hθ(p)θ(q);cis called Θ-invariant if and only if for all θ ∈ Θ and p, q ∈ V0, c(p, q) = c(θ(p), θ(q)); r is called Θ-invariantif and only if θ∈Θ such thatθ(χi(V0)) =χj(V0) =⇒ri =rj

. 3) Denote by FΘ120,FΘ1203 and FΘ123 theR-subcones of Θ-invariant forms in F120,F1203 and F123,respectively.

There are one-to-one correspondences between Θ-invariant forms, opera- tors and conductance matrices.

Remark 2.9.1) For affine nested fractals Θ is usually the symmetry group generated by the reflections in the hyperplanes bisecting the line segments which connects points in V0 and shall be denoted by Gs. Denote, also Fs120 :=

FG12s0,Fs1203 :=FG12s03,Fs123:=FG123s .

2) For an affine nested fractal S :=

n

F,{χi}i=1,No

and r Gs-invariant there exists α > 0 and A0 ∈ Fs123 (unique up to multiplication by a positive constant) such that Λ(A0) =αA0 (Cf. [3, 8, 12]).

(6)

3. RENORMALIZATION OF GENERALIZED “abc” GASKETS

In this section, we prove the existence of eigenforms for the renormal- ization map of a generalized “abc” gasket. Then we derive a necessary and sufficient condition in terms of ri, i = 1, N for the existence and uniqueness of the “highly symmetric” irreducible eigenoperator for the affine Sierpinski gasket and effectively determine it.

First, we recall a very useful tool called the “shortcut procedure”.

3.1. THE “SHORTCUT” PROCEDURE We consider S =

n

F,{χi}i=1,No

a connected post critical finite self- similar structure and r := (r1, . . . , rn), with ri > 0; consider also the initial boundary V0 of S and V1.

ConsiderPk:={Ak1, . . . , Akp}a partition ofVk,k= 0,1 andv:Vk−→R, v|Ak

i =ct, i = 1, p. Define v : Pk −→ R, v(Aki) := v(x), x ∈ Aki, i = 1, p.

Consider H0 ∈ F120(V0), so H1 := Ψ(H) ∈ F120(V1); let cHk : Vk×Vk −→ R be the conductance matrix associated to Hk. Define cHk : Pk× Pk −→ R, cHk(Aki, Akj) := P

p∈Aki,q∈Akj

cHk(p, q) for i6=j and cHk := 0 for i=j. Then (see Lemma 17 from [11])

(3.1) Hk(v, v) = 1 2

l

X

i,j=1

(v(Aki)−v(Akj))2cHk(Aki, Akj) =:Hk(v, v) and Hk∈F120(Pk) for k= 0,1.

For E ∈F1203(V0)\{0},F ∈B(V0) withE+F ∈F1203(V0) and u:Vk −→

R, obviously (Fk+nEk)(u, u) % ∞ for Ek(u, u) > 0 and (Fk+nEk)(u, u) = Fk(u, u) forEk(u, u) = 0, that isu =ct. on Ek-connected components. So, we can define (Fk+∞Ek)(u, u) :=Fk(u, u), foru∈KerEk, (Fk+∞Ek)(u, u) :=∞, for u /∈ KerEk. We say that we shortcut the edges of Ek. In (3.1) we consider Hk = Fk +nEk and v ∈ KerEk and we obtain Hk(v, v) = Fk(v, v) = (Fk+

∞Ek)(v, v) =Hk(v, v), so “shortcutting the edges of Ek” “the energy does not modify”.

3.2. RENORMALIZATION OF “abc” SIERPINSKI GASKETS I. Let ∆z1z2z3 ⊂R2 an equilateral triangle and H := co{z1, z2, z3}. Let {χi}i=1,M M similitudes so that Hi := χi(H) touches Hi−1 and Hi+1 in just one point; consider M = a+b+c, a ≥ 1, b ≥ 1, c ≥ 1; the choice of the

(7)

similitudes is so that we should havec+ 1 trianglesHi on z1z2,a+ 1 triangles Hi on the edge z2z3 and b+ 1 onz3z1 (see Figure 1, witha= 4,b= 5,c= 3).

H2 touches HM if and only if a=b= c= 1 (the classical Sierpinski gasket).

We denotei= 1, j:=c+ 1, k:=c+a+ 1, byF the attractor of the iterated function system

n

R2,{χi}i=1,Mo

and byχi also, the restrictions ofχi toF.

Fig. 1. The construction of a “453” gasket.

ThenSabc:=n

F,{χi}i=1,Mo

is a self-similar structure. B= S

1≤s<t≤M

(Hs∩Ht) hasM points and each ramifying point is coded by two addresses; for example, H1∩H2={π(1˙i) =π(2 ˙j)}, whereπ is the natural map and ˙l=ll . . .. So Γ has 2M codes, P ={(˙i),( ˙j),( ˙k)}andV0 ={z1, z2, z3}. Sabcis a post critical finite self-similar structure and is calledgeneralized “abc” gasket. For certainki and ratios we get an affine nested fractal with respect to the “maximal” groupGs. We take Θ ={IdV0}(IdV0 the identity map onV0) andr:= (r1, . . . , rM), with ri >0. For simplicity we consider equal weights for the similitudes with χi(H) on the same edge of the initial triangle (for example, if a = 4, b = 5, c= 3, then r2 =r3 =r4 =: 1/η2,r5 =r6 =r7 =r8 =: 1/η3,r9 =r10=r11= r12=r1 =: 1/η1, where η123 >0).

Because Θ ={IdV0}, the renormalization map will be considered on the largest cone. The coneF120 :=FΘ120 can be determined by looking at the eigen- values of an operatorH=H(α1, α2, α3) =

−α2−α3 α3 α2

α3 −α1−α3 α1 α2 α1 −α1−α2

,

(8)

α123≥0. Cf. [9].

F120 '

1, α2, α3)∈R3

α123 ≥0, α1α22α33α1≥0 , F1203 'R3+ and the F1203-parts are F(1,1,1)1203 '(0,∞)3,F(0,1,1)1203 ' {0} ×(0,∞)2, F(1,0,1)1203 ' (0,∞)× {0} ×(0,∞), F(1,1,0)1203 ' (0,∞)2 × {0}, F(1,0,0)1203 ' (0,∞)× {0} × {0},F(0,1,0)1203 ' {0} ×(0,∞)× {0},F(0,0,1)1203 ' {0} × {0} ×(0,∞).

Using Lemma 2.5 we get: Λ(F(1,0,0)1203 ) ⊂F(1,0,0)1203 (in Figure 2, just z2 and z3 can be connected by a path in Γ(Ψ(F(1,0,0)1203 )) with points from V1\V0) and similarly Λ(F(0,1,0)1203 ) ⊂ F(0,1,0)1203 , Λ(F(0,0,1)1203 ) ⊂ F(0,0,1)1203 and Λ(F(1,1,1)1203 ) ⊂ F(1,1,1)1203 . It can be shown that the only Λr-invariant F1203-parts are F1203 ' (0,∞)3, F1 := F(1,0,0)1203 ' (0,∞) × {0} × {0}, F2 := F(0,1,0)1203 ' {0} × (0,∞) × {0}, F3 :=F(0,0,1)1203 ' {0} × {0} ×(0,∞).

Fig. 2. Λ(D(1,0,0))D(1,0,0).

Now, we can state the following

Proposition 3.1. a)The eigenvalues of{Fi}i=1,3 are λ1 =

1 η2 + a

η3 −1

, λ2 = b

η2 + 1 η1

−1

, λ3 = 1

η3 + c η1

−1

. b) The eigenvalues of

Λr F1203+∞Fi i=1,3, are λ01 =

 1 η1

+

"

η2 b +

c−1 η1

+ 1 η3

−1#−1

−1

,

(9)

λ02 =

 1 η3 +

"

η1 c +

a−1 η3 + 1

η2

−1#−1

−1

,

λ03 =

 1 η2

+

"

η3 a +

b−1 η2

+ 1 η1

−1#−1

−1

.

If λi < λ0i, i= 1,3, then there exists irreducible eigenforms for Λr associated to Sabc.

Proof. a) Applying successively Ohm’s law (resistors in series add their resistances and resistors in parallel add their conductances) we getλi.

b) By “shortcutting” (3.1) and Ohm’s law we get the above λ0i.

BecauseF1203is Λ-invariant and for each Λ-invariantF1203-partFi ⊂∂F120

the eigenvalueλ0iis strictly greater thanλi, then Theorem 25 and Corollary 28 from [11] allow us to say that there exists an eigenform in F123.

II. Consider the particular case of “222” gasket; that is, we take M = 6, a = b = c = 2, and the scaling factors of the similitudes s1 = s2 = s3 = 2/5, s4 = s5 = s6 = 1/5. We take the “maximal” symmetry group Θ = Gs. If we denote by F the attractor of the iterated function system n

(R2,k · k);{χi, si}i=1,6o

and χi for their restrictions to F also, it is easy to verify thatS222 :=

n

F,{χi}i=1,6o

is an affine nested fractal with respect toGs and is calledthe affine Sierpinski gasket. The attractor is depicted in Figure 3.

Fig. 3. The attractor ofS222.

(10)

Consider r= (u, u, u, uv, uv, uv) (u,v >0) and D=

−2 1 1

1 −2 1

1 1 −2

.

D andr areGs-invariant and Γ(AD) = Γ((Fs1203)) is shown in Figure 4–(7).

Fig. 4. The renormalization of the affine Sierpinski gasket.

The graph Γ(Ψ((Fs1203))) is shown in (1). Using ∆ −Y and Y − ∆ transforms and Ohm’s law (Figure 4) we get

Proposition3.2. ForS222,Θ =Gsand the weightsr= (u, u, u, uv, uv, uv), the following statements are equivalent:

(1) there exists Gs invariant irreducible eigenoperators for the renormal- ization map Λr associated to S222.

(2) u(2v+ 5) = 3.

In this case,Dis the unique (up to a multiplication by a positive constant) eigenoperator of Λr for eigenvalue λ= u(2v+5)3 = 1.

So, in order to have existence and uniqueness even on this particular

“highly symmetric case”, the weights have to be considered with extreme care.

(11)

REFERENCES

[1] M. Barlow,Diffusions on Fractals. Lecture Notes in Mathematics, Springer, 1998.

[2] M. Barnsley,Fractals Everywhere. Academic Press, 1988.

[3] P.J. Fitzsimmons, B.M. Hambly and T. Kumagai,Transition density estimates for Brow- nian motion on affine nested fractals. Comm. Math. Phys. 165(1994), 595–620.

[4] J. Kigami,A harmonic calculus for p.c.f. self-similar sets. Trans. Amer. Math. Soc.

335(1993), 721–755.

[5] J. Kigami,Analysis on Fractals. Cambridge University Press, 2001.

[6] T. Kumagai, Regularity, closedness and spectral dimensions of the Dirichlet forms on p.c.f. self-similar sets. J. Math. Kyoto Univ. 33(1993), 765–786.

[7] S. Kusuoka,Diffusion processes on nested fractals. In: S. Kusuoka and R.L. Dobrushin (Eds.),Statistical Mechanics and Fractals. Springer: Berlin-Heidelberg, 1993.

[8] T. Lindstr¨om, Brownian motion on nested fractals. Mem. Amer. Math. Soc. 420 (1990).

[9] V. Metz, Hilbert projective metric on cones of Dirichlet forms. J. Funct. Anal. 127 (1995), 438–455.

[10] V. Metz, Renormalization contracts on nested fractals. J. Reine Angew. Math. 480 (1996), 161–175.

[11] V. Metz,The short-cut test. J. Funct. Anal. 220(2005), 118–156.

[12] C. Sabot,Existence and uniqueness of diffusions on finitely ramified self-similar fractals.

Ann. Sci. ´Ec. Norm. Sup´er. (4)30(1997), 605–673.

Received 10 October 2013 University of Pite¸sti,

Department of Mathematics and Informatics, Romania,

antonio 74nm@yahoo.com

Références

Documents relatifs

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des

We have assumed that the basic element of the generating pattem is defined by a single transfer function go(Tos), and we have shown that the limiting RTD is in dependent of go for

- We have studied, as a function of temperature and scattering angle, the intensity and time dependence of light scattered by the three-component liquid

For instance the Dupree renormalized dispersion relation for gyrokinetic, toroidal ion temperature gradient driven (ITG) modes, where the Dupree’s diffusion coefficient is assumed to

We state several results for the convergence of these orbits and give a geometrical description of the corresponding limit sets.. In particular, it provides a way to construct

This equation has the same form as the well-known Dirac- Hartree-Fock equations [40, 16], without exchange term, and with an addi- tional vacuum polarization potential induced by

In some way, a galactic space is a set with two levels of resolution: a fine resolution given by the generalized distance which relates the topological relations between points

in constructive theory, sharp cutoffs in direct or momentum space for ordinary field theory have bad decay properties in the dual Fourier variable, and this leads to difficulties