CC BY
3
Владикавказский математический журнал 2022, Том 24, Выпуск 4, С. 58-69
УДК 517.518.23+517.548.2 DOI 10.46698/w5793-5981-8894-o
ON POLETSKY-TYPE MODULUS INEQUALITIES FOR SOME CLASSES OF MAPPINGS*
S. K. Vodopyanov1
1 Sobolev Institute of Mathematics, 4 Akademika Koptyuga Ave., Novosibirsk 630090, Russia E-mail: vodopis@math.nsc.ru
Abstract. It is well-known that the theory of mappings with bounded distortion was laid by Yu. G. Reshetnyak in 60-th of the last century [1]. In papers [2, 3], there was introduced the two-index scale of mappings with weighted bounded (q,p)-distortion. This scale of mappings includes, in particular, mappings with bounded distortion mentioned above (under q = p = n and the trivial weight function). In paper [4], for the two-index scale of mappings with weighted bounded (q,p)-distortion, the Poletsky-type modulus inequality was proved under minimal regularity; many examples of mappings were given to which the results of [4] can be applied. In this paper we show how to apply results of [4] to one such class. Another goal of this paper is to exhibit a new class of mappings in which Poletsky-type modulus inequalities is valid. To this end, for n = 2, we extend the validity of the assertions in [4] to the limiting exponents of summability: 1 < q < p < to. This generalization contains, as a special case, the results of recently published papers. As a consequence of our results, we also obtain estimates for the change in capacity of condensers.
Key words: quasiconformal analysis, Sobolev space, modulus of a family of curves, modulus estimate. AMS Subject Classification: 30C65 (26B35, 31B15, 46E35).
For citation: Vodopyanov, S. K. On Poletsky-type Modulus Inequalities for Some Classes of Mappings, Vladikavkaz Math. J., 2022, vol. 24, no. 4, pp. 58-69. DOI: 10.46698/w5793-5981-8894-o.
1. Introduction
The goal of this work is to show the application of results of [4] for output of Poletsky-type modulus inequalities for some classes of mappings. For doing this we formulate first the main result of [4], and then we provide how it can be applied for some concrete classes of mappings. The main classes of mappings studied in [4] were defined in [2, 3].
Definition 1. Let w: Rn ^ [0, to] be a measurable function, called a weight, with 0 < w < to holding Hn-almost everywhere, and Q C Rn is a domain in Rn. A mapping f : Q ^ Rn with n ^ 2 is called a mapping with (inner) bounded w-weighted (q,p)-codistortion, or briefly, f € ID(Q; q,p; w, 1), where n — 1 ^ q ^ p < to, whenever
(1) f is continuous, open and discrete;
(2) f belongs to the Sobolev class Wn\-1 ioc(Q);
(3) the Jacobian determinant satisfies det Df (x) ^ 0 for almost all x € Q;
# The study was carried out within the framework of the State contract of the Sobolev Institute of Mathematics, project № FWNF-2022-0006. © 2022 Vodopyanov, S. K.
(4) the mapping f has bounded codistortion: adj Df (x) = 0 a. e. on the set Z = {x € Q : det Df (x) = 0};
(5) the local u-weighted (q,p)-codistortion function
n-l
O. ^^ n ) U 9 {x)ladi^ix)l if detU/C^^O,
U 3 X ^ Jtiq f(x, f) = \ det Df(x) P (1)
otherwise,
belongs to Lg(Q), where g satisfies ^ = ^^ — while g = oo for q = p.
Put K,p1(f;Q) = HK^O.f) I
Definition 2. Let u: Rn ^ [0, to] be a measurable function, called a weight, with 0 < u < to holding -almost everywhere, and Q C Rn is a domain in Rn. A mapping f : Q ^ Rn with n ^ 2 is called a mapping with (outer) bounded u-weighted (q,p)-distortion, or briefly f € OD (Q; q,p; u, 1), with n — 1 ^ q ^ p < to, whenever:
(1)f is continuous, open and discrete;
(2) f belongs to the Sobolev class W?1_1 ioc(Q);
(3) the Jacobian determinant satisfies det Df (x) ^ 0 for a. e. x € Q;
(4) the mapping f has bounded distortion: Df (x) = 0 a. e. on the set Z = {x € Q : det Df (x) = 0};
(5) the local u-weighted (q,p)-distortion function
J(x)\Df(x)\
if det Df (x) = 0,
n3x^K^(x,f) = <| det£>/(*)* ^^ , (2)
otherwise,
belongs to LJQ), where x satisfies - = \ — while x = oo for q = p.
X q p
Put Kqwp1(f; Q) = |KqWp1(-,f) I Lk (Q)||. Remark 1. It is established in [3] that
OD(Q; q,p; u, 1) C ID(Q; q,p; u, 1) (3)
in case of n — 1 <q ^ p < to.
For justifying (3) we refer to [3, Theorem 8] where it is proved that every mapping f : Q ^ Q' of OD(Q; q,p; u, 1), n — 1 < q ^ p < to, belongs also to the class ID(Q; q,p; u, 1), and the estimate
IlK^OJ) I Le(Q)|| < HK^OJ) I L^f-1 (4)
holds. (Here g and k are defined after formulas (1) and (2) respectively). In [4] it was proved the following result.
Theorem 1 [4, Theorem 4.1]. Let n — 1 <q ^ p< to. Suppose that f : Q ^ Rn is a mapping with with inner bounded u-weighted (q,p)-codistortion (f € ID(Q; q,p; u, 1)),
n-l
while the weight function 0(x) = u «-(«-1) (x) is locally summable. IfT is a family of curves in the domain Q then we have the inequality
(mods f (r))1/s < Kqp1 (f; Q) (mod? r)1/r, (5)
with s = „ P -in and r =
p—(to—1) g—(to— 1) '
q
Below we recall the concept of the modulus of a family of curves (see [4] for more details).
A curve in Rn is a continuous mapping a: I ^ Rn, where I is an interval in R, that is, a set of the form (a, b), where each angular parenthesis can be either round or square, a, b € R with a ^ b. We also allow infinite intervals. A curve a is called closed (open) if the interval I is compact (open). Put |a| = a(1). The expression y' c y will mean that the curve y' is a restriction of the curve y to a subinterval or a point.
If a: I = [a, b] ^ Rn is a closed curve then its length is
i
|a(
i= 1
1(a) = sup£ |a(ti) - a(ii+i)!
where the supremum is taken over all finite partitions a = t1 ^ t2 ^ ... ^ tl ^ tl+1 = b. If a curve a is not closed then put its length equal to 1(a) = sup l(a| j), where the supremum is taken over all closed subintervals J of I.
A curve a: I ^ Rn is called rectifiable whenever 1(a) < to. A curve is called locally rectifiable if each closed subcurve of it is rectifiable.
Consider a closed curve a: [a, b] ^ Rn and suppose that it is rectifiable. Define a function sa: [a, b] ^ R by the equality sa(t) = l(a|[a,t]). For the rectifiable curve a there exists a unique curve a0: [0,l(a)] ^ Rn obtained from a by a monotonely increasing change of parameter such that sao(t) = t and a = a0 o sa [5, Section 2.4]. The curve a0 is called the positive natural parametrization of a.
Take a Borel set A c Rn and a Borel function p: A ^ [0, to]. The integral of p along a rectifiable curve a: [a, b] ^ Rn is defined as
£(a)
„0/
j pds = J p(a°(r)) dH 1(t)
a
0
with an usual Lebesgue integral in the right-hand side. If a is absolutely continuous then so is the function sa(t) = [a, b] ^ [0,l(a)]. Putting t = sa(t) in the last integral, using the change-of-variables theorem for Lebesgue integrals, and accounting for a(t) = ^a0(sQ,(i))sQ,(i) and ■£pa°(T) = 1, we infer that
b
J pds = j p(a(t))|a(t)| dH 1(t). (6)
a a
Observe that by the change of variable formula we can express this as
b
J pds = j p(a(t))|a(t)| dH 1(t) = y p(y)N (y,a, [a,b]) dH1(y), (7)
a a |a|
where N(y, a, [a, b]) = #{[a, b] n a-1(y)} is the Banach indicatrix. For a locally rectifiable curve a: I ^ Rn, put
J pds = sup J pds, (8)
a ^ £
where the supremum is taken over all closed subcurves ^ of a.
Consider a family r of curves in Rn, where n ^ 2. A Borel function p : Rn ^ [0, to] is called admissible for r whenever
J pds ^ 1 (9)
Y
for each locally rectifiable curve y € r. Denote the collection of all admissible functions by adm r. Given a weight function d : Rn ^ (0, to) and a number p € [1, to), define the d-weighted p-modulus of r as
modP r= inf /pP0dHn.
p p€adm r J
Properties of the weight function will be prescribed separately; at least, we assume that it is locally summable and 0 < d < to holds Hn-almost everywhere. For d = 1 we obtain the usual definition of p-modulus, and instead of modi r we write modp r. If admT = 0 then we put modp r = to; this case is realized only if r contains at least one curve determining a constant mapping.
Remark 2. The definition of modulus implies that every family of curves which are not locally rectifiable has zero modulus. Moreover, if r is a family of curves and r = {7 € r : Y is locally rectifiable} then modp(r) = modP(ri).
Suppose that a is a rectifiable closed curve in Rn. A mapping g: |a| — Rn is called absolutely continuous on a if the composition g o a0 is absolutely continuous on [0,1(a)].
Theorem 2 [5, Fuglede's Theorem; 6]. Suppose that f: Q — Rn is a mapping of class Wi(Q) with 1 ^ p < to, and r is a family of locally rectifiable curves in Q such that each curve has a closed subcurve on which f is not absolutely continuous. Then modp r = 0.
2. Modification of Theorem 1 in the Case of n = 2 and p = to
In this case parameters q,p may be taken within (1, to]: 1 < q ^ p ^ to. The case 1 <q ^ p< to is taken into consideration in Theorem 1.
Theorem 3. Let 1 < q < p = to. Suppose that Q C R2 is a domain, and f: Q ^ R2
is a mapping with inner bounded w-weighted (q, to)-codistortion (f € ID(Q; q, to; w, 1)1),
__
while the weight function 0(x) = uj 1-1 (x) is locally summable. IfT is a family of curves in the domain Q then we have the inequality
(modi f (r)) < ;Q)(mod? r)1/r (10)
with r = ~r~r.
q—1
In this theorem Ks°o(f; Q) = ||Ks°o(-,f) I Lr(Q)||.
Theorem 4. Suppose that Q C R2 is a domain, and f: Q R2 is a mapping belonging to the Sobolev class W11loc(Q) with the nonnegative Jacobian determinant: det Df (x) ^ 0 for almost all x € Q. Assume that
1) f is continuous, open and discrete;
(2) the mapping f has bounded codistortion: adj Df (x) =0 a. e. on the set Z = {x € Q : det Df (x) = 0}.
In the case p = oc we have to replace det Df(x) ? in (1) by 1.
Let, for a weight u: Rn ^ [0, to], (to, TO)-codistortion function
Q 3 x ^ K-L(x,f) = (u(x)|adj Df (x)| lf det Df (x) > 0, (11)
0 otherwise,
belongs to ll(Q) (in another words f € ID(Q; to, to; u, 1)). If the weight function 0(x) = u-1(x) is locally summable then, for any family of curves r in the domain Q, we have the inequality
mod1 f (r) < (f ;Q)mod? r. (12)
In this theorem (f; Q) = (■, f) | L-(Q)||.
Theorems 3 and 4 will be proved in Section 6.
3. Application
In paper [7, Example 32] the following class of mappings is considered. Suppose that n — 1 < p < to, and consider a continuous, open and discrete mapping f : D' ^ Rn of an open connected domain D' C Rn, where n ^ 2, such that
(1) f € Wn_1;ioc (D');
(2) det Df (y) ^ 0 and f has finite codistortion; i. e., adj Df (y) = 0 Hn-almost everywhere on Z = {y € D' : det Df (y) = 0};
(3) the inner operator distortion function
D' 3 y ^ K„_\Mf ) =
|adjD/fa)|
n — 1
det Df(y) —
if det Df (y) = 0, otherwise,
(13)
belongs to LPt\oc(D'), where | = j- — ^^ holds with s = > n — 1;
(4) the weight function a defined as
f I adj Df{y)\v = J detD/(y)P-i
if y € D'\Z', otherwise,
(14)
is in € L1,ioc (D'), here Z' = {y € D' : Df (y) = 0}.
Taking into acount saying above we see that f : D' ^ D meets the assumptions of Theorem 1 with D' instead of Q: (2a) f € Wn_1;ioc (D');
(2b) det Df (y) ^ 0 and f has finite codistortion;
(2c) f: D' ^ D is a mapping of bounded u-weighted (s, s)-codistortion with u(y) = __
<t p~1(y), that is, the w-weighted (s, s)-codistortion function
D' 3 y ^ «'W) =
^(y)\ad}Df(y)\
, n —1
det Df(y)~
if J(y, f) = 0,
otherwise,
belongs to Ll(D') and
|KSTs,1(-,f) I l-(D')|| = 1
0
0
(the last equality is proved in [7, Theorem 3] under more general assumption).
Taking into account saying above, by Theorem 1, we come to the following statement.
Proposition 1. Suppose that a continuous, open and discrete mapping f : D' ^ Rn of an open connected domain D' c Rn, where n ^ 2, has the following properties:
(1) f e WUioc (D');
(2) det Df (y) ^ 0 and f has finite codistortion (adj Df (y) = 0 Hn-almost everywhere on Z = {y e D' : det Df (y) = 0});
(3) the inner operator distortion function
\ |adJD/(^i if det Df(y) / 0, D' 3 y X^L\s{y, /) = { detD/^-r- (16)
I 0 otherwise,
belongs to LPi\oc(Dr) with some p > n — 1, where | = — ^ir holds with s = > n — 1.
If r is a family of curves in the domain D' then we have the inequality
modp f (r) < modp r (17)
where the weight function a is defined in (7).
< When deriving inequality (17) the properties (2a)-(2c) formulated above, should be
taken into account. Really, we see that f e ID(Q; q,p; w, 1) with q = p = s and w(y) = __
(7 p-i(y). Therefore, by Theorem 1, we get the inequality
mods' f (r))1/s' < K-'1(f; D')( modf' r)1/s'
with s' = ^^y (here ^/{f-D') = ||I ¿coPOID- Because of (15), s' = p and
n-1
6>(y) = u) (y) = cr(y) inequality (17) holds. >
Taking into account [2, Theorem 34] or [4, Theorem 5.2] and its proof we come to
Proposition 2. Suppose that for a continuous, open and discrete mapping f : D' ^ Rn of an open connected domain D' c Rn, where n ^ 2, conditions of Proposition 1 hold. If E = (A, C) is a condenser in Q, then the estimate holds: capp f (E) ^ capp E.
4. The Special Case of the Mappings Under Consideration: n = 2
In the case n = 2 we have the following modification of the results of the previous section. We have 1 < p < to and a continuous, open and discrete mapping f : D' ^ R2 of on open connected domains D' c R2 such that
(1) f e Wlioc (D');
(2) det Df (y) ^ 0 and f has finite codistortion; i. e., adj Df (y) = 0 H2-almost everywhere on Z = {y e D' : det Df (y) = 0};
(3) the inner operator distortion function
, |adJD/(& if detD/(y)/0, D' 9 y J^"/) = < detD/(y) P
'p_1 In if detDf{y) = 0,
belongs to Lp,loc(D').
(4) the weight function a defined as
C | adj Df(y)\p -f DM 7,
y 1 otherwise,
is in € Li,ioc (D'), here Z' = {y € D' : Df (y) = 0}.
It is not hard to see that the continuous, open and discrete mapping f : D' — R2 meets the assumptions of Proposition 1 under n = 2: (3a) f € Wlioc (D'); (3b) f has finite distortion;
(3c) f:D'—ïD is a mapping with bounded w-weighted (j/,^-distortion where p' = __
and w(y) = a v^1{y)1 that is the w-weighted (p', ^-distortion function
, J if det Df(y) / 0,
D 3 y M- (y, /) = < det Df(y)V
0 otherwise,
belongs to LM(D'), and
||K&(f) I L-(D')|| = 1. (19)
Taking into account saying above, by Proposition 1, we come to the following statement.
Corollary 1. Suppose that a continuous, open and discrete mapping f : D' — R2 of an open connected domain D' C R2 has the following properties:
(1) f € W^ioc (D');
(2) f has finite codistortion (adj Df (y) = 0 H2-almost everywhere on Z = {y € D' : det Df (y) = 0});
(3) the inner operator distortion function
|adJD/fa)l if det Df(y) /0,
f\ — J .
ay
D' J^y, /) = <( detD/(y)7 (20)
0 otherwise,
belongs to LPy\oc(D') with some p > 1, where ^ + ^7 = 1-
If r is a family of curves in the domain D' then we have the inequality
modp f (r) < modp r (21)
holds where the weight function a is defined in (18).
5. One More Special Case of the Mappings Under Consideration: n = 2 and p = 1
In this section we prove that Corollary 1 is valid also in the case p = 1. To show this we have to modify some arguments of the previous section. A counterpart of Corollary 1 is formulated in the following statement.
Proposition 3. Suppose that a continuous, open and discrete mapping f : D' — R2 of an open connected domain D' C R2 has the following properties:
(1) f € W^ioc (D');
(2) det Df (y) ^ 0 and f has finite codistortion (adj Df (y) = 0 H2-almost everywhere on Z = {y e D' | det Df (y) = 0});
(3) the inner operator codistortion function
D ^ y ^ K 1,1 (y f) / |adj Df (y)| if det Df (y) = 0, D3 y — K1 (yf . (22)
0 otherwise,
belongs to L1loc(D').
If r is a family of curves in D' then we have
mod1 f (r) < mod^ r (23)
with a defined in (25).
< We show that the proof of Proposition 3 can be reduced to Theorem 3. For doing this formulate first additional properties of f and ^ = f-1.
Properties of ^ = f-1. If f : D' — D is a homeomorphism then the inverse homeomorphism ^ = f-1 : D — D' enjoys the following properties:
(4) by [9, Theorem 4] or [7, Theorem 27] we have ^ e W11loc(D) (see also [10, Theorem 3.2]);
(5) has finite distortion by [7, Theorem 27] (see also [10, Theorem 3.3]);
(6) is differentiable a. e. in D by [7, Theorem 27]; while f : D' — D
(6) belongs to Q1;1 (D,D'; a) (see [4]), that is the distortion function
Kl'Ux, = [ -(^»dit %x) if det D<p(x) / 0, 1,1 v [0 if det Dp(x)=0,
of the inverse mapping ^ = f-1 with the weight function a e L1loc (D') defined as
a(y) = 11 adj (y)l 'fr°'\Z'' where Z' = {y e D : Df (y) = 0}, (25)
1 otherwise,
is in Lo(D) and KD) = UK| Lo(D)|| = 1 (see [4, Theorem 25, formulas (30) and (37); 8]).
Properties of f. Taking into account saying above, we see thatf : D' — D meets some additional properties:
(7) f e W1 ioc (D') and f is differentiable a.e.in D' by [7, Theorem 27];
(8) det Df (y) ^ 0 and f has finite distortion by [7, Theorem 27] (see also [10, Theorem 3.3]);
(9) f: D' — D is a mapping with bounded w-weighted (to, TO)-codistortion with the weight function w = a-1, that is the w-weighted (to, TO)-codistortion function
D * y - KSi(»,/) = (w(y)|adj B/(y)| =
0 otherwise,
belongs to L oo (D'), and
IIK&(-,f) I Lo(D')|| = UK1 ; r(.,¥>) | Lo(D)U = 1. (26)
Now it is evident that f enjoys the conditions of Theorem 3, and therefore (23) holds for f. >
6. Proof of Theorems 3 and 4
< We verify that the proof of Theorem 1 given in [4, Theorem 4.1] for mappings with bounded ^-weighted (q,p)-codistortion, where n — 1 < q ^ p< to, works also in the case 1 < q ^ p = to at n = 2. To do this we need properties of Poletsky function and Poletsky's Lemma in this case. We formulate and prove them below. >
1. Properties of Poletsky function. Take a continuous mapping f : Q — R2 and a domain D compactly embedded into Q, meaning that D is bounded and D C Q, written briefly as D d Q, and take y / f(dD). Denote by ^(y,f,D) the degree of f at y with respect to D. Say that f is sense-preserving whenever ^(y, f, D) > 0 for all domains D d Q and all points y € f (D)\f (<9D). For A C Q refer as the multiplicity function to R2 9 y — N(y, f, A) = ## {f-1(y) n A}. Moreover, put N(f, A) = sup^ N(y, f, A).
Suppose that f : Q — R2 is a continuous, open, and discrete mapping. A domain D d Q is called normal whenever f (dD) = df (D). A normal neighborhood of x € Q is a normal domain [/Cii such that IJ n /_1(/(a:)) = {x}. The quantity i(x,f) = /x(/(x),/, [/) is independent of the choice of a normal neighborhood U of x (see [11, Chapter II, §2] for instance) and is called the local index of f at x. A point x € Q is called a branch point of f whenever f is not a homeomorphism of any neighborhood of x. Denote the collection of all branch points of f by Bf. If D is a normal domain for a mapping f then ^(y, f, D) is independent of y € f (D). We will call this constant by ^(f, D).
In the following two lemmas we state propositions of interest in their own right. Both of them are applied in the proof of the main result of this section.
Lemma 1 [3, Lemma 10]. Assume that f : Q — R2 is a continuous, open and discrete mapping in W11ioc(Q) with finite distortion. Then for every open connected set U C Q the set {x € U\Bf : J(x, f) = 0} has positive measure.
< If, on the contrary, J(x, f) =0 a. e. on a connected set U C Q\Bf on which f is a homeomorphism then Df (x) = 0 a. e. on U because f has finite distortion. Then f is constant on U, and consequently, f cannot be open. >
Proposition 4. If f : Q — R2 is a continuous, open and discrete mapping in W11ioc(Q) with finite distortion, then f is differentiable a. e. on Q\Bf and sense-preserving.
< For a connected open set U C Q\Bf on which f is a homeomorphism, it is enough to apply the statement [9, Theorem 4] or [7, Theorem 27] twice. For the restriction f iu : U — f (U) it provides that the inverse homeomorphism (f |u)-1 : f (U) — U is in W11(f(U)), is of finite distortion, and is differentiable a.e. on f(U). Then applying [7, Theorem 27] to (f|u)-1 : f(U) — U we get similar properties to the given mapping f |u : U — f (U). By Lemma 1, det Df (x) ^ 0 and properties of degree we conclude that f is sense-preserving. >
Definition 3. For a sense-preserving, continuous, open and discrete mapping f : Q — R2 and a normal domain D d Q, define the Poletsky function go : V — R2 on V = f (D) [12] by putting
V 9 y — go (y)=A £ i(x,f )x, (27)
xef-1(y)nD
where A = ^(f, D).
The function of the form (27) was introduced by Poletsky in [12] for mappings with bounded distortion (p = q = n, w = 1). The next statement presents the properties of the Poletsky function for the classes of mappings under consideration.
Proposition 5 [2, 3]. Suppose that f : Q — R2 belongs to OD(Q; to, to; w, 1) (properties (4a)-(4c) hold). Then
(1) the function gD defined in (27) is continuous and belongs to ACL(V);
(2) DgD (y) =0 a.e. on Z' U £';
(3) Poletsky function gD defined in (27) is in W-^V); furthermore,
||DgD | L1(V)U < AUK-'^(■; f) | L„(D)\\ J a(x) dX.
D
We emphasize that the formulated statement is proved in [2, Theorem 18] for mappings f e ID(Q;p,p; w, 1), p e (1, to). The same proof works also in the case p = to at n = 2.
2. Poletsky's Lemma. Consider a continuous, open and discrete mapping f : Q — R2. Take a closed rectifiable curve 5 : /0 — Rn and a curve a : / — Q with f oa C 5. In particular, we have / C /0. If the function s^ : /0 — [0,1(5)] is constant on some interval J C /, then the mapping 5 is constant on J. In turn, since f is discrete, a is also constant on J. Consequently, there exists a unique mapping a* : s^(/) — Q satisfying a = a* o s^ |7. We can prove that a* is continuous and f o a* C 50. The curve a* is called an f-representative of a (with respect to 5 ) whenever 5 = f ◦ a. Suppose now that 5 = f ◦ a. The above arguments show that
f o a* = (f o a)0.
Therefore, the curve f oa* admits a positive natural parametrization, and hence it is Lipschitz. Thus we can integrate along this curve using (6) where (/ o a*) (t) | = 1 for Jf 1-almost all t e /.
The mapping f is called absolutely precontinuous on a provided that a* is absolutely continuous.
Lemma 2. Suppose that f : Q — R2 is a mapping of class ID(Q; to, to; w, 1). Consider a family r of curves in Q such that for every 7 e r the following holds: the curve f o 7 is locally rectifiable and 7 has a closed subcurve a on which f is not absolutely precontinuous. Then mod 1 f (r) = 0.
The formulated Lemma is proved in [4, Lemma 3.3] for mappings f e ID(Q;p,p; w, 1), p e (1, to). The same proof works also in the case p = to at n = 2.
In the proof of Lemma 2 we also need the following statement.
Lemma 3. Consider a homeomorphism ^ : Q — Q' of class ID(Q; q, to; 0,1), where Q, Q' C R2 and 1 < q ^ to. Then
(1) the inverse homeomorphism is ^>-1 e W^ loc(Q');
(2) ^>-1 has finite distortion: D^>-1 (y) = 0 almost everywhere on Z';
L1 , r
(3) K1 (-,^-1) € where
r=i^+î ifq< 00, 1-1 if q < 00,
if q = 00, [61-1 if q = to;
(4) if the weight function u is locally summable then the inverse homeomorphism induces, by the change-of-variable rule, the bounded operator
: L1(q; u) n W^ >loc ^ Lftfi').
We have the relations
IlK1;(,^-1 ) | Le(fi')|| = | Le(fi)||
and
(-^-1) | Le(Q')|| < ||^-1*|| < ||Ki;rw) | Le(Q')||, where 0q,^ is some constant.
< Properties (1) and (2) of ^ = f-1 were proved just after Proposition 3. Taking into account (1) and (2) Properties (3) and (4) can be proved by analogy with Theorem 9 of [2]. > Remark 3. By means of Theorems 3 and 4 for homeomorphisms ^ : Q — Q' of class ID(Q; q, to; 0,1), where Q, Q' C R2 and 1 < q ^ to, we can prove some more inequalities such that Vaisala inequality and the capacity inequality (see proofs in [4, Theorem 22] and [4, Theorem 28] respectively).
Remark 4. It is not hard to see that assumptions of Theorem 4 are weaker comparing with those in paper [13]. For instance, Theorem 1.3 of [13] is formulated under addition condition that the given mapping is closed. Therefore Theorem 4 with weaker assumptions contains the main result of paper [13].
Acknowledgements. I greatly appreciate the anonymous reviewers for critically reading and comments, which helped improve the initial manuscript.
References
1. Reshetnyak Yu. G. Space Mappings with Bounded Distortion, Providence, Amer. Math. Soc., 1989.
2. Vodopyanov, S. K. Basics of the Quasiconformal Analysis of a Two-Index Scale of Space Mappings, Siberian Mathematical Journal, 2018, vol. 59, no. 5, pp. 805-834. DOI: 10.1134/S0037446618050075.
3. Vodopyanov, S. K. Differentiability of Mappings of the Sobolev Space W^1_1 with Conditions on the Distortion Function, Siberian Mathematical Journal, 2018, vol. 59, no. 6, pp. 983-1005. DOI: 10.1134/S0037446618060034.
4. Vodopyanov, S. K. Moduli Inequalities for W?1_1,loc-Mappings with Weighted Bounded (q,p)-Dis-tortion, Complex Variables and Elliptic Equations, 2021, vol. 66, no. 6-7, pp. 1037-1072. DOI: 10.1080/17476933.2020.1825396.
5. Vaisala J. Lectures on n-Dimensional Quasiconformal Mappings, Lecture Notes in Mathematics, vol. 229, Berlin-Heidelberg-New York, Springer, 1971.
6. Fuglede, B. Extremal Length and Functional Completion, Acta Mathematica, 1957, vol. 98, pp. 171-219. DOI: 10.1007/BF02404474.
7. Vodopyanov, S. K. The Regularity of Inverses to Sobolev Mappings and the Theory of Qq,p-Homeomorphisms, Siberian Mathematical Journal, 2020, vol. 61, no. 6, pp. 1002-1038. DOI: 10.1134/S0037446620060051.
8. Vodopyanov, S. K. and Tomilov, A. O. Functional and Analytic Properties of a Class of Mappings in Quasi-Conformal Analysis, Izvestiya: Mathematics, 2021, vol. 85, no. 5, pp. 883-931. DOI: 10.1070/IM9082.
9. Vodopyanov, S. K. Regularity of Mappings Inverse to Sobolev Mappings, Sbornik: Mathematics, 2012, vol. 203, no. 10, pp. 1383-1410. DOI: 10.1070/SM2012v203n10ABEH004269.
10. Hencl, S. and Koskela, P. Regularity of the Inverse of a Planar Sobolev Homeomorphism, Archive for Rational Mechanics and Analysis, 2006, vol. 180, pp. 75-95. DOI: 10.1007/s00205-005-0394-1.
11. Rickman S. Quasiregular mappings, Berlin, Springer-Verlag, 1993, 213 p.
12. Poletsky, E. A. The Modulus Method for Nonhomeomorphic Quasiconformal Mappings, Mathematics of the USSR-Sbornik, 1970, vol. 12, no. 2, pp. 260-270. DOI: 10.1070/SM1970v012n02ABEH000921.
13. Salimov, R. R., Sevost'yanov, E. A. and Targonskii, V. A. On Modulus Inequality of the Order p for the Inner Dilatation. arXiv - MATH - Complex Variables. 2022. DOI:arxiv-2204.07870.
Received September 2, 2022
Sergey K. Vodopyanov Sobolev Institute of Mathematics,
4 Akademika Koptyuga Ave., Novosibirsk 630090, Russia, Principal Re,searcher E-mail: vodopis@math.nsc.ru https://orcid.org/0000-0003-1238-4956
Владикавказский математический журнал 2022, Том 24, Выпуск 4, С. 58-69
О МОДУЛЬНЫХ НЕРАВЕНСТВАХ ТИПА ПОЛЕЦКОГО ДЛЯ НЕКОТОРЫХ КЛАССОВ ОТОБРАЖЕНИЙ
Водопьянов С. К.1
1 Институт математики им. С. Л. Соболева, Россия, 630090, Новосибирск, пр-т Академика Коптюга, 4 E-mail: vodopis@math.nsc.ru
Аннотация. Хорошо известно, что теория отображений с ограниченным искажение была заложена Ю. Г. Решетняком в 60-е годы прошлого века [1]. В работах [2, 3] была введена двухиндексная шкала отображений с весовым ограниченным (д,р)-искажением. Эта шкала отображений включает в себя, в частности, отображения с ограниченным искажением, упомянутые выше (при q = p = n и тривиальной весовой функции). В работе [4] для двухиндексной шкалы отображений с весовым ограниченным ^,р)-искажения доказано модульное неравенство типа Полецкого при минимальной регулярности; приведено много примеров отображений, к которым можно применить результаты [4]. В этой статье мы приведем одно такое применение. Другая цель этой статьи — показать новый класс отображений, в которых выполняются модульные неравенства типа Полецкого. Для этого мы расширяем при n = 2 справедливость утверждений работы [4] на предельные показатели: 1 < q < p < то. Это обобщение содержит в качестве частного случая результаты недавно опубликованных работ. Как следствие результатов этой статьи мы получаем также оценки изменения емкости конденсаторов.
Ключевые слова: квазиконформный анализ, пространство Соболева, модуль семейства кривых, оценка модуля.
AMS Subject Classification: 30C65 (26B35, 31B15, 46E35).
Образец цитирования: Vodopyanov S. K. On Poletsky-Type Modulus Inequalities for Some Classes of Mappings // Владикавк. мат. журн.—2022.—Т. 24, № 4.—C. 58-69 (in English). DOI: 10.46698/w5793-5981-8894-o.
