Show Summary Details
More options …

Editor-in-Chief: Radulescu, Vicentiu / Squassina, Marco

IMPACT FACTOR 2018: 6.636

CiteScore 2018: 5.03

SCImago Journal Rank (SJR) 2018: 3.215
Source Normalized Impact per Paper (SNIP) 2018: 3.225

Mathematical Citation Quotient (MCQ) 2018: 3.18

Open Access
Online
ISSN
2191-950X
See all formats and pricing
More options …
Volume 9, Issue 1

# Regularity for minimizers for functionals of double phase with variable exponents

Maria Alessandra Ragusa
• Corresponding author
• Dipartimento di Matematica e Informatica, Viale Andrea Doria, 6-95125, Catania, Italy
• RUDN University”, 6 Miklukho - Maklay St, Moscow, 117198, Russia
• Email
• Other articles by this author:
/ Atsushi Tachikawa
• Department of Mathematics, Faculty of Science and Technology, Tokyo University of Science, Noda, Chiba, 278-8510, Japan
• Email
• Other articles by this author:
Published Online: 2019-07-20 | DOI: https://doi.org/10.1515/anona-2020-0022

## Abstract

The functionals of double phase type

$H(u):=∫|Du|p+a(x)|Du|qdx, (q>p>1, a(x)≥0)$

are introduced in the epoch-making paper by Colombo-Mingione [1] for constants p and q, and investigated by them and Baroni. They obtained sharp regularity results for minimizers of such functionals. In this paper we treat the case that the exponents are functions of x and partly generalize their regularity results.

## 1 Introduction and main theorem

The main goal of this paper is to provide a regularity theorem for minimizers of a class of integral functionals of the calculus of variations called of double phase type with variable exponents defined for uW1,1(Ω; ℝN) (Ω ∈ ℝn, n, N ≥ 2) as

$F(u,Ω):=∫Ω|Du|p(x)+a(x)|Du|q(x)dx, q(x)≥p(x)>1, a(x)≥0,$

where p(x), q(x) and a(x) are assumed to be Hölder continuous. They do not only have strongly non-uniform ellipticity but also discontinuity of growth order at points where a(x) = 0. The above functional is provided by the following type of functionals with variable exponent growth

$u↦∫g(x,Du)dx, λ|z|p(x)≤g(x,z)≤Λ(1+|z|)p(x), Λ≥λ>0,$

which are called of p(x)-growth. These p(x)-growth functionals have been introduced by Zhikov [2] (in this article α(x) is used as variable exponents) in the setting of Homogenization theory. He showed higher integrability for minimizers and, on the other hand, he gave an example of discontinuous exponent p(x) for which the Lavrentiev phenomenon occurs ([3, 4]).

Such functionals provide a useful prototype for describing the behaviour of strongly inhomogeneous materials whose strengthening properties, connected to the exponent dominating the growth of the gradient variable, significantly change with the point. In [3], Zhikov pointed out the relationship between p(x)-growth functionals and some physical problems including thermistor. As another application, the theory of electrorheological materials and fluids is known. About these objects see, for example, [5, 6, 7, 8].

These kind of functionals have been the object of intensive investigation over the last years, starting with the inspiring papers by Marcellini [9, 10, 11], where he introduced so-called (p, q)- or nonstandard growth functionals:

$u↦∫f(x,u,Du)dx, λ|z|p≤f(x,u,z)≤Λ(1+|z|)q, q≥p≥1, Λ≥λ>0.$

About general (p, q)-growth functionals, see for example [3, 4, 12, 13, 14, 15, 16, 17, 18, 19] and the survey [20].

For the continuous variable exponent case, nowadays many results on the regularity for minimizer are known, see [21, 22, 23, 24]. Further results in this direction can be, for instance, found in [25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41] for partial regularity results for p(x)-energy type functionals:

$u↦∫Aijαβ(x,u)Dαui(x)Dβuj(x)p(x)dx, Aijαβ(x,u)zαizβj≥λ|z|2$

In 2015 a new class of functional so-called functionals of double phase are introduced by Colombo-Mingione [1]. In the primary model they have in mind are

$u↦H(u;Ω):=∫H(x,Du)dx, H(x,z):=|z|p+a(x)|z|q,$

where p and q are constants with qp > 1 and a(⋅) is a Hölder continuous non-negative function. By Colombo-Mingione [1, 42, 43] and Baroni-Colombo-Mingione [44, 45, 46] many sharp results are given about the regularity of local minimizers of the functional defined as

$u↦G(u;Ω):=∫ΩG(x,u,Du)dx,$(1.1)

where G(x, u, z) : Ω × ℝ × ℝnR is a Carathéodory function satisfying the following growth condition for some constants Λλ > 0 besides several natural assumptions:

$λH(x,z)≤G(x,u,z)≤ΛH(x,z).$

For the scalar valued case, in [46] regularity results are given comprehensively. Under the conditions

$a(⋅)∈C0,α(Ω), α∈(0,1] and qp≤1+αn,$(1.2)

or

$u∈L∞(Ω), a(⋅)∈C0,α(Ω), α∈(0,1] and qp≤1+αp,$(1.3)

they showed that a local minimizer of 𝓖 defined as (1.1) is in the class C1,β for some β ∈ (0, 1).

For the scaler valued case, see also [47]. They proved Harnack’s inequality and the Hölde continuity for quasiminimizer of the functional fo type

$∫φ(x,|Du|)dx,$

where φ is the so-called Φ-function. We mention that Harnack’s inequality is not valid in the vector valued cases which we are considering in the present paper.

On the other hand, for vector valued case, in [1], under the condition

$a(⋅)∈C0,α(Ω), α∈(0,1] and qp<1+αn,$(1.4)

C1,β-regularity, for some β ∈ (0, 1), of local minimizers is given.

Zhikov has given in [3, 4] examples of functionals with discontinuous growth order for which Lavrentiev phenomenon occurs. So, in general settings, we can not expect regularity of minimizers for such functionals which change their growth order discontinuously. So, conditions (1.2), (1.3) and (1.4), which guarantee the regularity of minimizers, are very significant.

In this paper we deal with a typical type of functionals of double phase with variable exponents and show a regularity result for minimizers.

In our opinion these results present new and interesting features from the point of view of regularity theory.

Let Ω ⊂ ℝn be a bounded domain, p(x), q(x) and a(x) functions on Ω satisfying

$p,q∈C0,σ(Ω), q(x)≥p(x)≥p0>1, forall x∈Ω$(1.5)

where p0 is a fixed constant strictly larger than one and

$a∈C0,α(Ω), a(x)≥0,$(1.6)

for α, σ ∈ (0, 1]. Moreover, we assume that p(x) and q(x) satisfy

$supx∈Ω q(x)p(x)<1+βn, β=min{α,σ},$(1.7)

at every xΩ (compare these conditions with (1.2)). Let F : Ω × ℝnN → [0, ∞) be a function defined by

$F(x,z):=|z|p(x)+a(x)|z|q(x).$(1.8)

We consider the functional with double phase and variable exponents defined for u : Ω → ℝN and DΩ as

$F(u,D)=∫DF(x,Du)dx.$(1.9)

For a bounded open set Ω ⊂ ℝn and a function p : Ω → [1, +∞), we define Lp(x)(Ω; ℝN) and W1,p(x)(Ω; ℝN) as follows:

$Lp(x)(Ω;RN):={u∈L1(Ω;RN) ; ∫Ω|u|p(x)dx<+∞}.W1,p(x)(Ω;RN):={u∈Lp(x)∩W1,1(Ω;RN) ; Du∈Lp(x)(Ω;RnN)}.$

In what follows we omit the target space ℝN. We also define $\begin{array}{}{L}_{\mathrm{l}\mathrm{o}\mathrm{c}}^{p\left(x\right)}\end{array}$(Ω) and $\begin{array}{}{W}_{\mathrm{l}\mathrm{o}\mathrm{c}}^{1,p\left(x\right)}\end{array}$(Ω) similarly. As mentioned in [48], if p(x) is uniformly continuous and ∂Ω satisfies uniform cone property, then

$W1,p(x)(Ω)={u∈W1,1(Ω) ;Du∈Lp(x)(Ω)}.$

Let us define local minimizers of 𝓕 as follows:

#### Definition 1.1

A function uW1,1Ω) is called to be a local minimizer of 𝓕 if F(x, Du) ∈ L1(Ω) and satisfies

$F(u;suppφ)≤F(u+φ;suppφ),$

for any φ$\begin{array}{}{W}_{\mathrm{l}\mathrm{o}\mathrm{c}}^{1,p\left(x\right)}\end{array}$(Ω) with compact support in Ω.

The main result of this paper is the following:

#### Theorem 1.2

Assume that the conditions (1.5), (1.6) and (1.7) are fulfilled. Let uW1,1(Ω) be a local minimizer of 𝓕. Then u$\begin{array}{}{C}_{\mathrm{l}\mathrm{o}\mathrm{c}}^{1,y}\end{array}$(Ω) for some y ∈ (0, 1).

#### Remark 1.3

(About the symbols for Hölder spaces) If we follow the standard textbooks, Dacorogna [49], Evans [50], Gilberg-Trudinger [51], etc., for k ∈ ℕ, 0 < α ≤ 1, Ck,α(Ω) mean the subspaces of Ck(Ω) consisting of functions whose k-th order partial derivatives are locally Hölder continuous. However, recently many authors (especially ones who study regularity problems) write them as $\begin{array}{}{C}_{\mathrm{l}\mathrm{o}\mathrm{c}}^{k,\alpha }\end{array}$(Ω), and they use Ck,α(Ω) for Ck,α(Ω̄) (namely, for uniformly Hölder continuous cases). Anyway, with “loc" there is no doubt of misunderstanding. So, in this paper we follow their usage for Hölder spaces.

In order to prove the above theorem, we employ a freezing argument; namely we consider a frozen functional which is given by freezing the exponents, and compare a minimizer of the original functional under consideration with that of frozen one.

## 2 Preliminary results

In what follows, we use C as generic constants, which may change from line to line, but does not depend on the crucial quantities. When we need to specify a constant, we use small letter c with index.

For double phase functional with constant exponents, namely for

$H(u,D):=∫DH(x,Du)dx, H(x,z)=|z|p+a(x)|z|q,$(2.1)

we prepare the following Sobolev-Poincaré inequality which is a slightly generalised version of [1, Theorem 1.6] due to Colombo-Mingione.

#### Theorem 2.1

Let a(x) ∈ C0,β(Ω) for some β ∈ (0, 1) and 1 < p < q constants satisfying

$qp<1+βn,$

and let ωL(ℝn) with ω ≥ 0 andBR ω dx = 1 for BRΩ with R ∈ (0, 1). Then, there exists a constant C depending only on n, p, q, [a]0,β, RnωL andDwLp(BR) and exponents d1 > 1 > d2 depending only on n, p, q, β such that

$∫−BRHx,u−〈u〉ωRd1dx1d1≤C∫−BRHx,Dud2dx1d2$(2.2)

holds whenever uW1,p(BR), where

$〈u〉ω:=∫BRu(x)ω(x)dx.$

Note that for the special choice ω = |BR|–1 χBR we have

$〈u〉ω=∫−BRu(x)dx.$

#### Proof

We can proceed exactly as in the proof of [1, Theorem 1.6] only replacing (3.11) of [1] by

$|u(x)−〈u〉ω|R≤CR∫BR|Du(y)||x−y|n−1dy,$

which is shown by [52, Lemma 1.50] (see also the proof of [53, Theorem 7]).□

From the above theorem, we have the following corollary.

#### Corollary 2.2

Assume that all conditions of Theorem 2.1 are satisfied, and let D be a subset of BR with positive measure. Then, there exists a constant C depending only on n, p, q, [a]0,β, Rn/|D| andDuLp(BR) and exponents d1 > 1 > d2 depending only on n, p, q, β such that the following inequality holds whenever uW1,p(x)(BR) satifies u ≡ 0 on D:

$∫−BRHx,uRd1dx1d1≤C∫−BRHx,Dud2dx1d2.$(2.3)

#### Proof

Choosing ω so that

$ω(x)=0x∈BR∖D1|D|x∈D$

and applying Theorem 2.1, we get the assertion.□

#### Remark 2.3

In [1, Theorem 6.1], and therefore also in the above theorem and corollary, the exponent d2 ∈ (0, 1) is chosen so that the following conditions hold:

$qp<1+βd2n$(2.4)

$pq(n−1)+1>1d2.$(2.5)

In fact, in [1], they choose a constant y ∈ (1, p) so that

$qp<1+αyn and p+q(n−1)yq(n−1)>1,$

(see [1, (3.6), (3.14)]), and put d2 = 1/y. Let us mention the that if d2 satisfies (2.4) and (2.5) for some q = q0 and p = p0, then the same d2 satisfies these inequalities for any q and p with q/pq0/p0.

For any yΩ and R > 0 with BR(x) ⊂ Ω let us put

$p2(y,R):=supBR(y)p(x),p1(y,R):=infBR(y)p(x),$(2.6)

$q2(y,R):=supBR(y)q(x),q1(y,R):=infBR(y)q(x).$(2.7)

We prove interior higher integrability of the gradient of a minimizer, similar results are contained in [54].

#### Proposition 2.4

Let u$\begin{array}{}{W}_{\mathrm{l}\mathrm{o}\mathrm{c}}^{1,p\left(x\right)}\end{array}$(Ω) be a local minimizer of 𝓕. Then, for any compact subset KΩ, F(x, Du) ∈ L1+δ0(K) and there exists a positive constant δ0 and C depending only on the given data and K such that

$∫−BR/2(y)F(x,Du)1+δ0dx11+δ0≤C+C∫−BR(y)F(x,Du)dx$(2.8)

holds for any BR(y) ⋐ K.

#### Proof

Let KΩ be a compact subset and R0 ∈ (0, dist(K, ∂Ω)) a constant such that

$0(2.9)

For any $\begin{array}{}{x}_{0}\in \stackrel{\circ }{K},\end{array}$ put

$κ0:=141+βn−supx∈BR(x0)q(x)p(x)>0.$(2.10)

Then, letting xR0(x0) be a such that p(x) = p1(x0, R0), we have

$q2(x0,R0)p1(x0,R0)=q(x−)+q2(x0,R0)−q(x−)p1(x0,R0)≤supx∈BR0(x0)q(x)p(x)+2σ[q]0,σR0σp0≤supx∈BR0(x0)q(x)p(x)+121+βn−supx∈BR0(x0)q(x)p(x)=121+βn+supx∈BR0(x0)q(x)p(x)≤1+βn−2κ0$(2.11)

The above estimate (2.11) implies that

$q2(x0,R0)<(p1(x0,R0))∗=np1(x0,R0)n−p1(x0,R0).$(2.12)

For any BR(y) ⊂ BR0(x0) with 0 < R < 1, and 0 < tsR, let η be a cut-off function such that η ≡ 1 on Bt(y), η ≡ 0 outside Bs(y) and $\begin{array}{}|D\eta |\le \frac{2}{s-t}.\end{array}$ Put w := uη(uuR), where uR = $\begin{array}{}{\int \phantom{\rule{-9pt}{0ex}}-}_{{B}_{R}\left(y\right)}\end{array}$ udx. Since

$Dw=(1−η)Du+(u−uR)Dη,$

we have

$F(x,Dw)≤c0[((1−η)|Du|)p(x)+|u−uR||Dη|p(x)+a(x)((1−η)|Du|)q(x)+|u−uR||Dη|q(x)],$

where c0 is a constant depending only on maxK q(x). On the other hand, since F(x, Du) ∈ L1, we have

$u∈W1,p(x)⊂W1,p1(x0,R0)⊂Lp1(x0,R0)∗⊂Lp2(x0,R0)⊂Lq(x),$

on BR0(x0). Thus, mentioning also that w = u outside Bs(y), we see that F(x, Dw) ∈ L1 (K), namely w is an admissible function. In the following part of the proof, let us abbreviate

$pi:=pi(y,R), qi:=qi(y,R) (i=1,2).$

Then, we have

$∫Bs(y)F(x,Du)dx≤∫Bs(y)F(x,Dw)dx≤ c0∫Bs(y)(1−η)p(x)(|Du|p(x)+a(x)|Du|q(x))dx +c0∫Bs(y)u−uRs−tp(x)+a(x)u−uRs−tq(x)dx≤ c0∫Bs(y)∖Bt(y)F(x,Du)dx+c0(s−t)p2∫Bs(y)|u−uR|p(x) +c0(s−t)q2∫Bs(y)a(x)|u−uR|q(x)dx$(2.13)

We can use hole-filling method. Add c0Bs(y)∖Bt(y) F(x, Du) dx to the both side and divide them by c0 + 1, then we get

$∫Bt(y)F(x,Du)dx≤c0c0+1∫Bs(y)F(x,Du)dx+1(s−t)p2∫Bs(y)|u−uR|p(x)dx+1(s−t)q2∫Bs(y)a(x)|u−uR|q(x)dx.$(2.14)

Using an iteration lemma [55, Lemma 6.1], we see, for some constant C = C(c0, p2, q2), that

$∫Bt(y)F(x,Du)dx≤C(s−t)p2∫Bs(y)|u−uR|p(x)+C(s−t)q2∫Bs(y)a(x)|u−uR|q(x)dx.$

Putting s = R and t = R/2, we have

$∫BR2(y)F(x,Du)dx≤CRp2∫BR(y)|u−uR|p(x)+CRq2∫BR(y)a(x)|u−uR|q(x)dx ≤CRp1−p2∫BR(y)u−uRRp(x)dx+CRq1−q2∫BR(y)a(x)u−uRRq(x)dx ≤CRp1−p2∫BR(y)1+u−uRRp2dx+CRq1−q2∫BR(y)1+a(x)1q(x)u−uRRq2dx.$(2.15)

Since Rp1p2 and Rq1q2 are bounded because of the Hölder continuity of exponents p(x) and q(x), putting

$a~(x):=a(x)q2q(x),$

from (2.15), we obtain the estimate

$∫BR2(y)F(x,Du)dx≤CRn+CRn∫−BR(y)u−uRRp2dx+a~(x)u−uRRq2dx=:I+II.$(2.16)

In order to get the boundedness of Rp1p2 and Rq1q2 the so-called “log-Hölder continuity" (see [56, section 4.1]) is sufficient. On the other hand by virtue of the Hölder continuity of q(⋅), we have that ãC0,β (β = min{α, σ}). Let d2 ∈ (0, 1) be a constant satisfying (2.4) and (2.5) for β = min{α, σ}, q = q2(x0, R0) and p = p1(x0, R0). Then, for any BR(y) ⊂ BR0(x0), this d2 satisfy (2.4) and (2.5) with q = q2(y, R) and p = p2(y, R).

By Theorem 2.1, we can estimate II as follows.

$II≤CRn∫−BR(y)|Du|p2+a~(x)|Du|q2d2dx1d2≤CRn∫−BR(y)|Du|d2p2dx1d2+CRn∫−BR(y)a(x)1q(x)|Du|d2q2dx1d2.$(2.17)

As mentioned above, (2.17) holds for for any BR(y) ⊂ BR0(x0) with same d2. Now, take R > 0 sufficiently small so that

$d2p2(y,R)

and let θ ∈ (d2, 1) be a constant satisfying

$d2p2(y,R)<θp1(y,R) and d2q2(y,R)<θq1(y,R).$(2.18)

Then, using Hölder inequality, we can estimate the first term of the right hand side of (2.17) as follows.

$∫−BR(y)|Du|d2p2dx1d2≤∫−BR(y)|Du|θp1dxp2θp1= ∫−BR(y)|Du|θp1dxp2−p1θp1⋅∫−BR(y)|Du|θp1dx1θ≤ ∫−BR(y)(1+|Du|p(x))dxp2−p1θp1⋅∫−BR(y)1+|Du|θp1dx1θ.$(2.19)

Since,

$∫BR(y)|Du|p(x)dx≤F(u,BR(y))≤F(u,K)$

and u locally minimizes 𝓕, ∫BR(y) |Du|p(x) dx is bounded. On the other hand, as mentioned after (2.15), R–(p2p1) is bounded. So, there exists a constant c1 = c1 (𝓕(u, K), p(x), d2, n, θ)

$∫−BR(y)|Du|p(x)dxp2−p1θp1≤(ωnRn)−(p2−p1)θp1F(u,K)p2−p1θp1≤c1(F(u,K),p(x),d2,n,θ),$

where ωn denotes the volume of a n-dimensional unit ball. Thus, from (2.19) we obtain for some positive constant c2 = c2(c1, θ)

$∫−BR(y)|Du|d2p2dx1d2≤c2+c2∫−BR(y)|Du|θp(x)dx1θ.$(2.20)

Similarly, we can estimate the second term of the left hand side of (2.17) as follows.

$∫−BR(y)a(x)1q(x)|Du|d2q2dx1d2≤∫−BR(y)a(x)1q(x)|Du|θq1dxq2θq1≤∫−BR(y)a(x)1q(x)|Du|θq1dxq2−q1θq1∫−BR(y)a(x)1q(x)|Du|θq1dx1θ≤∫−BR(y)1+a(x)1q(x)|Du|q(x)dxq2−q1θq1∫−BR(y)1+a(x)1q(x)|Du|θq(x)dx1θ.$(2.21)

As above, using local minimality of u and the fact that R–(q2q1) is bounded, we have for a positive constant c3 = c3(𝓕(u, K), q(x), d2, n, θ)

$∫−BR(y)a(x)1q(x)|Du|d2q2dxq2−q1θq1≤c3(F(u,K),q(x),d2,n,θ).$(2.22)

Thus, we obtain for some positive constant c4 = c4(c3, θ)

$∫−BR(y)a(x)1q(x)|Du|d2q2dx1d2≤c4+c4∫−BR(y)a(x)1q(x)|Du|θq(x)dx1θ.$(2.23)

Combining (2.16), (2.17), (2.20) and (2.23), we see that there exists a constant C depending on the given data and 𝓕(u, K) such that

$∫−BR2(y)F(x,Du)dx≤C+C∫−BR(y)F(x,Du)θdx1θ$(2.24)

for any BR(y) ⊂ BR0KΩ. Now, by virtue of the reverse Hölder inequality with increasing domain due to Giaquinta-Modica [57], we get the assertion.□

For δ0 determined in Proposition 2.4, in what follows, we always take R > 0 sufficiently small so that

$1+δ02p2(y,R)≤(1+δ0)p1(y,R) and 1+δ02q2(y,R)≤(1+δ0)q1(y,R).$(2.25)

We need also higher integrability results on the neighborhood of the boundary. Let us use the following notation: for T > 0 we put

$BT:=BT(0), BT+:={x∈Rn ; |x|0},ΓT:={x∈Rn ; |x|

We say “f = g on ΓT" when for any η$\begin{array}{}{C}_{0}^{\mathrm{\infty }}\end{array}$(BT) we have (fg)η$\begin{array}{}{W}_{0}^{1,1}\left({B}_{T}^{+}\right).\end{array}$ For yBT, we write

$Ωr:=Br(y)∩BT+.$

Then, we have the following proposition on the higher integrability near the boundary, independently proved in [58, Lemma 5], see also [59, Lemma 5] for the manifold constrained case.

#### Proposition 2.5

Let a(x), q and p satisfy the same conditions in Theorem 2.1 and let for A$\begin{array}{}{B}_{T}^{+}\end{array}$

$H(w,A):=∫AH(x,w)dx, H(x,z):=|z|p+a(x)|z|q.$

uW1,p($\begin{array}{}{B}_{T}^{+}\end{array}$) be a given function with

$∫BT+|Du|p+a(x)|Du|q1+δ0dx<∞,$

for some δ0 >. Assume that vW1,p (B+(T)) be a local minimizer of 𝓗 in the class

${w∈W1,p(BT+) ; u=w on ΓT}$

Then, for any S ∈ (0, T), there exists a constants δ ∈ (0, δ0) and C > 0 such that for any y$\begin{array}{}{B}_{S}^{+}\end{array}$ and R ∈ (0, TS) we have

$∫−ΩR/2H(x,Dv)1+δdx11+δ≤C∫−ΩRH(x,Dv)dx+C∫−ΩRH(x,Du)1+δdx11+δ.$

#### Proof

For convenience, we extend u, v, Du, Dv to be zero in BT$\begin{array}{}{B}_{T}^{+}\end{array}$. Of course, because extended u, v may have discontinuity on ΓT, they are not always in $\begin{array}{}{W}_{\mathrm{l}\mathrm{o}\mathrm{c}}^{1,p}\end{array}$ (BT), and therefore Du, Dv do not necessarily coincide with distributional derivatives of u, v on B(T). On the other hand, since u = v on Γ(T), uv is in the class W1,p(B(S)) and DuDv can be regarded as the weak derivatives of uv on B(S) for any S < T.

Let R be a positive constant satisfying R ≤ (TS)/2. For x0$\begin{array}{}{B}_{S}^{+}\end{array}$, we treat the two cases $\begin{array}{}{x}_{0}^{n}\le \frac{3}{4}R\end{array}$ and $\begin{array}{}{x}_{0}^{n}>\frac{3}{4}R\end{array}$ separately.

• Case 1

Suppose that $\begin{array}{}{x}_{0}^{n}\le \frac{3}{4}R\end{array}$. Take radii s, t so that 0 < R/2 ≤ t < sR and choose a η$\begin{array}{}{C}_{0}^{\mathrm{\infty }}\end{array}$(BT) such that 0 ≤ η ≤ 1, η ≡ 1 on Bt, supp ηBs and || ≤ 2/(st). Defining

$φ:=η(v−u),$

we see that $\begin{array}{}\phi \in {W}_{0}^{1,1}\left({B}_{T}^{+}\right)\end{array}$ with supp φBs, and that

$D(v−φ)=(1−η)Dv−(v−u)Dη+ηDu.$

Then, by virtue of the minimality of v, for a positive constant c4 depending only on q, we have

$∫ΩtH(x,Dv)dx≤∫ΩsH(x,Dv)dx≤∫ΩsH(x,D(v−φ))dx=∫Ωs|D(v−φ)|p+a(x)|D(v−φ)|qdx≤c4∫Ωs∖Ωt|Dv|p+a(x)|Dv|qdx+c4∫Ωs|Du|p+a(x)|Du|qdx+c4∫Ωs2s−tp|v−u|p+a(x)2s−tq|v−u|qdx≤c4∫Ωs∖Ωt|Dv|p+a(x)|Dv|qdx+c4∫Ωs|Du|p+a(x)|Du|qdx+c42s−tp∫Ωs|v−u|pdx+c42s−tq∫Ωsa(x)|v−u|qdx.$

Now, we use the hole filling method as in the proof of Proposition 2.4. Namely, adding

$c4∫Ωt|Dv|p+a(x)|Dv|qdx$

and dividing both side by c4 + 1, we obtain

$∫ΩtH(x,Dv)dx≤c4c4+1∫ΩsH(x,Dv)dx+∫ΩsH(x,Du)dx+1(s−t)p∫Ωs|v−u|pdx+1(s−t)q∫Ωsa(x)|v−u|qdx,$

Using the iteration lemma [55, Lemma 6.1], we get for some constant C = C(c4, p, q)

$∫ΩtH(x,Dv)dx≤C∫ΩsH(x,Du)dx+C(s−t)p∫Ωs|v−u|pdx+C(s−t)q∫Ωsa(x)|v−u|qdx.$

Putting t = R/2 and s = R, we have

$∫ΩR/2H(x,Dv)dx≤C∫ΩRHx,v−uRdx+C∫ΩRH(x,Du)dx.$

Let us now consider the mean integral in all the terms, we obtain

$∫−ΩR/2H(x,Dv)dx≤C∫−ΩRH(x,Du)dx+C∫−ΩRHx,v−uRdx.$

Since we are assuming that $\begin{array}{}{x}_{0}^{n}\le \frac{3}{4}R\end{array}$ we can apply Corollary 2.2 with a constant independent on R for the last term in the right hand side and get

$∫−ΩR/2H(x,Dv)dx≤C∫−ΩRH(x,Du)dx+C∫−ΩR(H(x,D(v−u)))d2dx1d2.$

Taking into consideration that d2 < 1 we share in the last term Dv and Du, apply Hölder inequality for the integral of H(x, Du)d2, and obtain

$∫−ΩR/2H(x,Dv)dx≤C∫−ΩRH(x,Du)dx+C∫−ΩRH(x,Dv)d2dx1d2.$(2.26)

• Case 2

Let us deal with the case that $\begin{array}{}{x}_{0}^{n}>\frac{3}{4}R\end{array}$. In this case, since B3R/4(x0) ⋐ $\begin{array}{}{B}_{T}^{+}\end{array}$, we can proceed as in [1, 9. Proof of Theorem 1.1:(1.8)], slightly modifying the radii, to get

$∫−ΩR/2H(x,Dv)dx=∫−BR/2H(x,Dv)dx≤C∫−B3R/4H(x,Dv)d2dx1d2≤C′∫−ΩRH(x,Dv)d2dx1d2.$(2.27)

Thus, we see that (2.26) holds for every 0 < R < (ST)/2. Now, the reverse Hölder inequality allows us to obtain

$∫−ΩRH(x,Dv)1+δdx11+δ≤C∫−ΩR2H(x,Dv)dx+C∫−ΩRH(x,Du)1+δdx11+δ.$

By virtue of [1, Theorem 1.1] and Proposition 2.5, we have the following global higher integrability for functions which minimize 𝓗 with Dirichlet boundary condition.

#### Corollary 2.6

Let a(x), q and p satisfy the same conditions in Theorem 2.1 and δ2 ∈ (0, 1) be a some constant. Assume that uW1,(1+δ1)p(BR(y)) be a given function with

$∫BR(y)H(x,Du)1+δ1dx:=∫BR(y)|Du|p+a(x)|Dv|q1+δ1dx≤C$

for some constant C > 0. Let vW1,p(BR(y)) be a minimizer of

$H(w,BR(y):=∫BR(y)H(x,Dw)dx$

in the class

$u+W01,p(BR(y))={w∈W1,p(BR(y)) ; u−w∈W01,p(BR(x0))}.$

Then, for some δ2 ∈ (0, δ1) and for any δ3 ∈ (0, δ2), we have H(x, Dv) ∈ L1+δ(BR(y)) and

$∫BRH(x,Dv)1+δ3dx≤C∫BRH(x,Du)1+δ3dx.$(2.28)

#### Proof

From [1, Theorem 1.1], Proposition 2.5 and covering argument, we have

$∫−BRH(x,Dv)1+δdx11+δ≤C∫−BRH(x,Dv)dx+C∫−BRH(x,Du)1+δdx11+δ$

and then, by the minimality of v,

$∫−BRH(x,Dv)1+δdx11+δ≤C∫−BRH(x,Du)dx+C∫−BRH(x,Du)1+δdx11+δ$

Once again we use the Hölder inequality for the first term of the right-hand side that gives us the assertion.□

## 3 Proof of the main theorem

In this section we prove Theorem 1.2. We employ the so-called direct approach, namely we consider a frozen functional for which the regularity theory has been established in [1] and compare a local minimizer of the frozen functional with u under consideration.

For a constant p > 1, let us define the auxiliary vector field Vp : ℝn → ℝn as

$Vp(z):=|z|p−2z.$(3.1)

Let mention that Vp satisfies

$|Vp(z)|2=|z|p and |Vp(z1)−Vp(z2)|≈(|z1|+|z2|)p−22|z1−z2|.$(3.2)

Proof of Theorem 1.2. We divide the proof into two parts. We prove the Hölder continuity of u in Part 1, and of the gradient Du in Part 2.

Part 1. Let K and BR0(x0), are as in the Proposition 2.4. For BR(y) ⊂ B2R(y) ⊂ BR0(x0), let us define pi and qi as in the Proposition 2.4. We define a frozen functional 𝓕0 as

$F0(x,z):=|z|p2+a(x)q2q(x)|z|q2$(3.3)

$F0(w,D)=∫BR(y)F0(x,Dw)dx.$(3.4)

In what follows, let us abbreviate $\begin{array}{}\stackrel{~}{a}\left(x\right)={\left(a\left(x\right)\right)}^{\frac{{q}_{2}}{q\left(x\right)}}\end{array}$ as in the proof of Proposition 2.4.

Let vWp2(BR(y)) be a minimizer of 𝓕0 in the class

$u+W0p2(BR(y)):={w∈Wp2(BR(y)) ; w−u∈W0p2(BR(y))}.$

Then, by [1, Theorem1.3], for any y ∈ (0, 1) there exists a constant C > 0 dependent on n, p2, q2, λ, Λ, [ã]0,β, ∥ã, ∥DvLp2(BR(y)) and y such that

$∫Bρ(y)F0(x,Dv)dx≤CρRn−y∫BR(y)F0(x,Dv)dx≤CρRn−y∫BR(y)F0(x,Du)dx,$(3.5)

where we used the minimality of v. Here, we mention that by the coercivity of the functional and the minimality of v we have the following:

$∥Dv∥Lp2(BR(y))p2≤F0(v,BR(y))≤F0(u,BR(y)).$(3.6)

On the other hand, since we are taking R > 0 sufficiently small so that (2.25) holds, there exists a constant C(p2, q2) > 0 such that

$F0(x,ξ)≤C(p2,q2)(1+F(x,ξ))1+δ0$(3.7)

holds for any (x, ξ) ∈ BR(y) × ℝnN. Now, by virtue of above 2 estimates and Proposition 2.4, we can see, for a constant C > 0 depending only on the given data on the functional, that

$∥Dv∥Lp2(BR(y))p2≤F0(v,BR(y))≤C1+F(u,K)1+δ.$(3.8)

Because of the local minimality of u, the last quantity is finite. Consequently, we can regard the constant in (3.5) is a constant depending only on given data and 𝓕(u, K).

For further convenience, let us mention that from (3.5), is nothing to see that

$∫Bρ(y)(1+F0(x,Dv))dx≤CρRn−y∫BR(y)(1+F0(x,Dv))dx≤CρRn−y∫BR(y)(1+F0(x,Du))dx.$(3.9)

Let us compare Du and Dv. Mentioning the elementary equality for a twice differentiable function

$f(1)−f(0)=f′(0)+∫01(1−t)f′′(t)dt,$

as [21, (9)], and using the fact that v satisfies the Euler-Lagrange equation of 𝓕0, we can see that

$F0(u)−F0(v)=∫BR(y)ddtF0(x,tDu−(1−t)Dv)|t=0dx+∫BR(y)dx∫01(1−t)d2dt2F0(x,tDu+(1−t)Dv)dt=∫BR(y)DzF0(x,Dv)(Du−Dv)+∫BR(y)dx∫01(1−t)DzDzF0(x,tDu+(1−t)Dv)(Du−Dv)(Du−Dv)dt≥C∫BR(y)dx∫01(1−t)|tDu+(1−t)Dv|p2−2+a~(x)|tDu+(1−t)Dv|q2−2|Du−Dv|2dt≥C∫BR(y)|Du|p2−2+|Dv|p2−2|Du−Dv|2dx+∫BR(y)a~(x)|Du|q2−2+|Dv|q2−2|Du−Dv|2dx.$(3.10)

On the other hand, by the minimality of v, we have

$F0(u)−F0(v)≤F0(u)−F(u,BR(y))+F(v,BR(y))−F0(v).$(3.11)

Since we are assuming p(x), q(x) ∈ C0,σ, using the inequality [21, (7)], we can see that, for any ε ∈ (0, 1), there exists a positive constant C such that

$F0(u)−F(u,BR(y))≤∫BR(y)|Du|p2−|Du|p(x)+a(x)1q(x)|Du|q2−a(x)1q(x)|Du|q(x)dx≤C(ε)Rσ∫BR(y)1+|Du|(1+ε)p2dx +C(ε)Rσ∫BR(y)1+a(x)1q(x)|Du|(1+ε)q2dx≤CRn+σ+C(ε)Rσ∫BR(y)1+|Du|p2(1+ε)+1+a~(x)|Du|q21+εdx≤CRn+σ+C(ε)Rσ∫BR(y)F0(x,Du)1+εdx$(3.12)

Similarly we have

$F(v,BR(y))−F0(v)≤∫BR(y)|Dv|p2−|Dv|p(x)+a(x)1q(x)|Dv|q2−a(x)1q(x)|Dv|q(x)dx≤C(ε)Rσ∫BR(y)1+|Dv|(1+ε)p2dx +C(ε)Rσ∫BR(y)1+a(x)1q(x)|Dv|(1+ε)q2dx≤ CRn+σ+C(ε)Rσ∫BR(y)1+|Dv|p2(1+ε)+1+a~(x)|Dv|q21+εdx≤ CRn+σ+C(ε)Rσ∫BR(y)F0(x,Dv)1+εdx.$(3.13)

Now, for δ0 of Proposition 2.4, choose δ3 > 0 so that (2.28) of Corollary 2.6 holds, and let us take ε so that ε ∈ (0, min{δ0/2, δ3}/2). Since we are choosing R so that (2.25) holds, we have

$F0(x,⋅)1+ε≤(1+F0(x,⋅))1+min{δ0/2,δ3}≤C(1+F(x,⋅))1+δ0.$(3.14)

By Proposition 2.4 and (3.14), we deduce from (3.12) that

$F0(u)−F(u,BR(y))≤CRn+σ+C(ε)Rσ∫BR(y)1+F(x,Du)1+δ0dx≤CRn+σ+CRσ∫BR(y)F(x,Du)1+δ0dx≤CRn+σ+CRσ−nε∫B2R(y)F(x,Du)dx1+δ0≤CRn+σ+CRσ−nε∫B2R(y)F(x,Du)dx,$(3.15)

where we used the fact that

$∫B2R(y)F(x,Du)dx≤∫KF(x,Du)dx≤M0$

for some constant M0. The existence of M0 guaranteed by the local minimality of u.

For (3.13) we use Proposition 2.6, Proposition 2.4 and (3.14), to get

$F(v,BR(y))−F0(v)≤CRn+σ+C(ε)Rσ∫BR(y)F0(x,Du)1+εdx≤CRn+σ+CRσ−nε∫B2R(y)F(x,Du)dx.$(3.16)

On the other hand, by the definition of F0, we have

$F(x,Du)≤C1+F0(x,Du).$

So we have, combining (3.10), (3.11), (3.15) and (3.16), that

$∫BR(y)|Du|p2−2+|Dv|p2−2|Du−Dv|2dx+∫BR(y)a~(x)|Du|q2−2+|Dv|q2−2|Du−Dv|2dx≤F0(u)−F0(v)≤CRn+σ+CRσ−nε∫B2R(y)(1+F0(x,Du))dx.$(3.17)

By virtue of (3.2) and (3.9), we can see that

$∫Bρ(y)(1+F0(x,Du))dx=∫Bρ(y)(1+F0(x,Dv))dx+∫Bρ(y)F0(x,Du)−F0(x,Dv)dx≤CρRn−y∫B(y)(1+F0(x,Dv))dx+∫Bρ(y)|Vp2(Du)|2+a~(x)|Vq2(Du)|2−|Vp2(Dv)|2+a~(x)|Vq2(Dv)|2dx≤CρRn−y∫B(y)(1+F0(x,Dv))dx+∫BR(y)|Vp2(Du)|2−|Vp2(Dv)|2+a~(x)|Vq2(Du)|2−|Vq2(Dv)|2dx≤ CρRn−y∫B(y)(1+F0(x,Dv))dx+∫BR(y)|Vp2(Du)−Vp2(Dv)|2dx+∫BR(y)a~(x)|Vq2(Du)−Vq2(Dv)|2dx≤ CρRn−y∫B(y)(1+F0(x,Dv))dx+∫BR(y)|Du|p2−2+|Dv|p2−2|Du−Dv|2dx +∫BR(y)a~(x)|Du|q2−2+|Dv|q2−2|Du−Dv|2dx≤ CρRn−y∫BR(y)(1+F0(x,Dv))dx +CRn+σ+CRσ−nε∫B2R(y)(1+F0(x,Du))dx≤CρRn−y+Rσ−nε∫B2R(y)(1+F0(x,Du))dx+CRn+σ.$(3.18)

Using well-known lemma (see for example [1, Lemma 5.13]), for sufficiently small R > 0, we can see that for any y ∈ (y, 1) there exists a constant C depending given data and ζ such that

$∫Bρ(y)F0(x,Du)dx≤CρRn−y′∫B2R(y)F0(x,Du)dx+Cρn−y′$(3.19)

hold for any ρ ∈ (0, R). Now, since (3.9) holds for any y ∈ (0, 1), we can choose y ∈ (0, 1) arbitrarily in (3.19). On the other hand, since we are supposing that p(x) ≥ p0 > 1, for any ζ ∈ (0, 1), choosing y ∈ (0, 1) so that yp0(1 – ζ), we see that there exists a positive constant C dependent on the given data, KΩ and 𝓕(u, K) such that

$∫Bρ(y)|Du|p0dx≤Cρn−p0(1−ζ)$

holds for any Bρ(y) with 4ρ ≤ dist (K, Ω). So, we conclude that u$\begin{array}{}{C}_{\mathrm{l}\mathrm{o}\mathrm{c}}^{0,\zeta }\end{array}$(Ω) for any ζ ∈ (0, 1) by virtue of Morrey’s theorem.

Part 2. Now, we are going to show the Hölder continuity of the gradient Du. For y$\begin{array}{}\stackrel{\circ }{K}\end{array}$ let R1 ∈ (0, R0) be a constant such that BR1(y) ⊂ K, and for 0 < R < R1/4 let v be as in Part 1. Then, by the estimate given by Colombo-Mingione at [1, p.484, l.-6], we see that there exist constants C > 0, dependent on n, p2, q2, λ, Λ, ∥ ã, dist(K, ∂Ω), 𝓕0(v, BR(y)) and α̃ ∈ (0, 1)

$∫−Bρ(y)|Dv−(Dv)ρ|p2dx≤Cρα~β64n,$(3.20)

holds for any ρR/2. Here, as in Part 1, let us mention that 𝓕0(v, BR(y)) can be controlled by 𝓕(u, K) as (3.8). So, we can choose the above constant in (3.20) to be dependent only on the given data of the functional, the local minimizer u under consideration and K.

In what follows, let us abbreviate

$α¯:=α~β64n.$

By virtue of (3.20), for ρ and R as above, we get

$∫Bρ(y)|Du−(Du)ρ|p2dx≤C∫Bρ(y)|Du−(Dv)ρ|p2dx≤C∫Bρ(y)Dv−(Dv)ρp2dx+C∫Bρ(y)Du−Dvp2dx≤Cρn+α¯+C∫BR(y)Du−Dvp2dx.$(3.21)

For the case that p2 ≥ 2, since there exists a constant such that

$|z1−z2|p2≤C|z1|p2−2+|z2|p2−2|z1−z2|2$

for any z1, z2 ∈ ℝn, using (3.17), we can estimate the last term of the right hand side of (3.21) as

$∫BR(y)|Du−Dv|p2dx≤CRn+σ+CRσ−nε∫B2R(y)F0(x,Du)dx.$(3.22)

We use (3.19) replacing ρ by 2R and R by R0 to see that

$∫B2R(y)F0(x,Du)dx≤CRn−ζR0ζ∫−BR0F0(x,Du)dx+CRn−ζ.$

Since R0 is determined in the beginning of the proof, we can regard $\begin{array}{}{R}_{0}^{\zeta }\text{\hspace{0.17em}}{\int \phantom{\rule{-9pt}{0ex}}-}_{{B}_{{R}_{0}}}{F}_{0}\left(x,Du\right)dx\end{array}$ as a constant. So, we get

$∫B2R(y)F0(x,Du)dx≤CRn−ζ.$(3.23)

By (3.22) and (3.23), we obtain

$∫BR(y)|Du−Dv|p2dx≤CRn+σ+CRn−ζ+σ−nε≤CRn−ζ+σ−nε.$(3.24)

When 1 < p2 < 2, using Hölder’s inequality, (3.2) and (3.17), we can see that

$∫BR(y)|Du−Dv|p2dx≤C∫BR(y)Vp2(Du)−Vp2(Dv)p2(|Du|+|Dv|)p2(2−p2)2dx≤C∫BR(y)Vp2(Du)−Vp2(Dv)2dxp22∫BR(y)(|Du|+|Dv|)p22dx2−p22≤∫BR(y)(|Du|+|Dv|)p2−2|Du−Dv|2dxp2∫BR(y)F0(x,Du)dx2−p22≤CRn+σ+CRσ−nε∫B2R(y)F0(x,Du)dxp22∫B2R(y)F0(x,Du)dx2−p22≤CR(n+σ)p22∫B2R(y)F0(x,Du)dx2−p22+CR(σ−nε)p22∫B2R(y)F0(x,Du)dx.$(3.25)

By (3.25) and (3.23), we obtain

$∫BR(y)|Du−Dv|p2dx≤CRp2(n+σ)2R(2−p2)(n−ζ)2+CR(σ−nε)p22Rn−ζ=CRn−ζ+p2(σ+ζ)2+CRn−ζ+p2(σ−nε)2≤2CRn−ζ+p2(σ−nε)2≤2CRn−ζ+(σ−nε)2.$(3.26)

For the last inequality we used the following facts:

$01.$

Mentioning the above facts again and comparing (3.24) and (3.26), we see that, for p2 > 2, the estimate (3.26) holds. Now, combining (3.21) and (3.26), we obtain

$∫Bρ(y)|Du−(Du)ρ|p2dx≤Cρn+α¯+Rn−ζ+σ−nε2.$

This holds for any 0 < ρ < R/2 ≤ R0/8. For k > 1, let us put ρ = Rk/2 (bearing in mind that Rk/2 ≤ R/2 holds for k > 1), then

$ρn+α¯+Rn−ζ+σ−nε2=ρn+α¯+(2ρ)2n−2ζ+σ−nε2k.$

So, we have

$∫Bρ(y)|Du−(Du)ρ|p2dx≤ρn+α¯+(2ρ)2n−2ζ+σ−nε2k.$(3.27)

Since

$α¯=α~64nβ=α~64nmin{α,σ}≤σ64,$

we can take ε sufficiently small so that < (σ)/2 then, for sufficiently small ζ,

$n−ζ+σ−nε2>n+α¯$

holds. Now, for such a choice of ε and ζ, putting

$k=2n−2ζ+σ−nε2(n+α¯) (>1)$

in (3.27), we get

$∫Bρ(y)|Du−(Du)ρ|p2dx≤Cρn+α¯,$

and therefore we obtain the Hölder continuity of Du by virtue of the Campanato’s theorem.

## Acknowledgement

The authors are deeply grateful to Giuseppe Mingione for interesting them in the problem. This paper was partly prepared while the authors visited in Pisa the Centro di Ricerca Matematica Ennio De Giorgi – Scuola Normale Superiore in September 2016. The hospitality of the center is greatly acknowledged.

The first author is partially supported by PRIN 2017 and the Ministry of Education and Science of the Russian Federation (5-100 program of the Russian Ministry of Education). The second author is partially supported by Japan Society for the Promotion of Science KAKENHI Grant Number 17K05337.

## References

• [1]

M. Colombo and G. Mingione, Regularity for double phase variational problems. Arch. Ration. Mech. Anal., 215 (2), (2015), 443–496.

• [2]

V. V. Zhikov, Averaging of functionals of the calculus of variations and elasticity theory. Izv. Akad. Nauk SSSR Ser. Mat., 50 (4), (1986), 675–710. Google Scholar

• [3]

V. V. Zhikov, On Lavrentiev’s phenomenon. Russian J. Math. Phys., 3 (2), (1995), 249–269. Google Scholar

• [4]

V. V. Zhikov, On some variational problems. Russian J. Math. Phys., 5 (1997), 105–116. Google Scholar

• [5]

M. Ružička, Electrorheological fluids: modeling and mathematical theory, 1748 of Lecture Notes in Mathematics. Springer-Verlag, Berlin, 2000. Google Scholar

• [6]

E. Acerbi and G. Mingione, Regularity results for stationary electro-rheological fluids. Arch. Ration. Mech. Anal., 164 (3), (2002), 213–259.

• [7]

K. Rajagopal and M. Ružička, Mathematical modeling of electrorheological materials. Contin. Mech. Thermodyn., 13, (2001), 59–78.

• [8]

V. Bögelein, F. Duzaar, J. Habermann and C. Scheven, Stationary electro-rheological fluids: low order regularity for systems with discontinuous coefficients. Adv. Calc. Var., 5 (1), (2012), 1–57.

• [9]

P. Marcellini, Regularity of minimizers of integrals of the calculus of variations with nonstandard growth conditions. Arch. Rational Mech. Anal., 105 (3), (1989), 267–284.

• [10]

P. Marcellini, Regularity and existence of solutions of elliptic equations with p, q-growth conditions. J. Differential Equations, 90 (1), (1991), 1–30.

• [11]

P. Marcellini, Everywhere regularity for a class of elliptic systems without growth conditions. Ann. Scuola Norm. Sup. Pisa Cl. Sci. (4), 23 (1), (1996), 1–25. Google Scholar

• [12]

M. Bildhauer and M. Fuchs, C1,α-solutions to non-autonomous anisotropic variational problems. Calc. Var. Partial Differential Equations, 24 (3), (2005), 309–340.

• [13]

D. Breit, New regularity theorems for non-autonomous variational integrals with (p, q)-growth. Calc. Var. Partial Differential Equations, 44 (1-2), (2012), 101–129.

• [14]

H. J. Choe, Interior behaviour of minimizers for certain functionals with nonstandard growth. Nonlinear Anal., 19 (10), (1992),933–945.

• [15]

L. Esposito, F. Leonetti and G. Mingione, Sharp regularity for functionals with (p, q) growth. J. Differential Equations, 204 (1), (2004), 5–55.

• [16]

T. Schmidt, Regularity theorems for degenerate quasiconvex energies with (p, q)-growth. Adv. Calc. Var., 1 (3), (2008), 241–270. Google Scholar

• [17]

T. Schmidt, Regularity of minimizers of W1,p-quasiconvex variational integrals with (p, q)-growth. Calc. Var. Partial Differential Equations, 32 (1), (2008),1–24.

• [18]

T. Schmidt, Regularity of relaxed minimizers of quasiconvex variational integrals with (p, q)-growth. Arch. Ration. Mech. Anal., 193 (2), (2009),311–337.

• [19]

N. N. Ural′ tseva and A. B. Urdaletova, Boundedness of gradients of generalized solutions of degenerate nonuniformly elliptic quasilinear equations. Vestnik Leningrad. Univ. Mat. Mekh. Astronom., (vyp. 4), (1983), 50–56. Google Scholar

• [20]

G. Mingione, Regularity of minima: an invitation to the dark side of the calculus of variations. Appl. Math., 51 (4), (2006),355–426.

• [21]

A. Coscia and G. Mingione, Hölder continuity of the gradient of p(x)-harmonic mappings. C. R. Acad. Sci. Paris Sér. I Math., 328 (4), (1999), 363–368.

• [22]

E. Acerbi and G. Mingione, Regularity results for a class of quasiconvex functionals with nonstandard growth. Ann. Scuola Norm. Sup. Pisa Cl. Sci. (4), 30 (2), (2001), 311–339. Google Scholar

• [23]

E. Acerbi and G. Mingione, Regularity results for a class of functionals with non-standard growth. Arch. Ration. Mech. Anal., 156 (2), (2001), 121–140.

• [24]

M. Eleuteri, Hölder continuity results for a class of functionals with non-standard growth. Boll. Unione Mat. Ital. Sez. B Artic. Ric. Mat. (8), 7 (1), (2004), 129–157. Google Scholar

• [25]

M. Cencelj, V. Rădulescu and D. Repovš, Double phase problems with variable growth. Nonlinear Analysis, 177, (2018), 270–287.

• [26]

M. Eleuteri, P. Marcellini and E. Mascolo, Lipschitz estimates for systems with ellipticity conditions at infinity. Ann. Mat. Pura Appl., 195, (2016), 1575-1603.

• [27]

M. Eleuteri, P. Marcellini and E. Mascolo, Regularity for scalar integrals without structure conditions Adv. Calc. Var., in press. Google Scholar

• [28]

M. Eleuteri, P. Marcellini and E. Mascolo, Lipschitz continuity for energy integrals with variable exponents Rend. Lincei - Matematica e Appl., 27, (2016), 61–87.

• [29]

G. Mingione, Short tales from nonlinear calderón-Zygmund theory. Lecture Notes in Mathematics., 2186, (2017), 159–204.

• [30]

N.S. Papageorgiou, V. Rădulescu and D. Repovš, Double-phase problems with reaction of arbitrary growth Zeitschrift fuer angewandte Mathematik und Physik (ZAMP), 69, (2018), 108.

• [31]

V. Rădulescu and D. Repovš, Partial Differential Equations with Variable Exponents: Variational Methods and Qualitative Analysis CRC Press, Taylor Francis Group, Boca Raton FL, 2015. Google Scholar

• [32]

V. Rădulescu and Q. Zhang, Double phase anisotropic variational problems and combined effects of reaction and absorption terms J. Math. Pures Appl., 118, (2018), 159–203.

• [33]

V. Rădulescu, Nonlinear elliptic equations with variable exponent: old and new Nonlinear Analysis: Theory, Methods and Applications, 121, (2015), 336–369.

• [34]

V. Rădulescuu, Isotropic and anisotropic double-phase problems: old and new Opuscula Mathematica, 39, (2019), 259–279.

• [35]

V. Rădulescu, D. Repovš, X. Shi and Q. Zhang, Multiple solutions of double phase variational problems with variable exponent Advances in Calculus of Variations, , in press.

• [36]

M. A. Ragusa, A. Tachikawa and H. Takabayashi, Partial regularity of p(x)-harmonic maps. Trans. Amer. Math. Soc., 365 (6), (2013), 3329–3353. Google Scholar

• [37]

M. A. Ragusa and A. Tachikawa, On interior regularity of minimizers of p(x)-energy functionals. Nonlinear Anal., 93, (2013), 162–167.

• [38]

A. Tachikawa, On the singular set of minimizers of p(x)-energies. Calc. Var. Partial Differential Equations, 50 (1-2), (2014), 145–169.

• [39]

K. Usuba, Partial regularity of minimizers of p(x)-growth functionals with 1 < p(x) < 2. Bull. Lond. Math. Soc., 47 (3), (2015), 455–472.

• [40]

M. A. Ragusa and A. Tachikawa, Boundary regularity of minimizers of p(x)-energy functionals. Ann. Inst. H. Poincaré Anal. Non Linéaire, 33 (2), (2016), 451–476.

• [41]

A. Tachikawa and K. Usuba, Regularity results up to the boundary for minimizers of p(x)-energy with p(x) > 1. Manuscripta Math., 152 (1-2), (2017), 127–151.

• [42]

M. Colombo and G. Mingione, Bounded minimisers of double phase variational integrals. Arch. Ration. Mech. Anal., 218 (1), ( 2015), 219–273.

• [43]

M. Colombo and G. Mingione, Calderón-Zygmund estimates and non-uniformly elliptic operators. J. Funct. Anal., 270 (4), (2016), 1416–1478.

• [44]

P. Baroni, M. Colombo and G. Mingione, Nonautonomous functionals, borderline cases and related function classes. St.Petersburg Math. J., 27 (3), ( 2016), 347–379.

• [45]

P. Baroni, M. Colombo, and G. Mingione, Harnack inequalities for double phase functionals. Nonlinear Anal., 121, (2015), 206–222.

• [46]

P. Baroni, M. Colombo, and G. Mingione, Regularity for general functionals with double phase. Calc. Var. Partial Differential Equations, 57 (2), (2018), art. 62, 48. Google Scholar

• [47]

P. Harjulehto, P. Hästö and O. Toivanen, Hölder regularity of quasiminimizers under generalized growth conditions, Calc. Var. Partial Differential Equations, 56, (2017), no. 2, article 22. Google Scholar

• [48]

A. Coscia and D. Mucci, Integral representation and Γ-convergence of variational integrals with P(X)-growth. ESAIM Control Optim. Calc. Var., 7, (2002), 495–519 (electronic).

• [49]

B. Dacorogna, Introduction to the calculus of variations. Imperial College Press, London, (2015), third edition. Google Scholar

• [50]

L. C. Evans, Partial differential equations, volume 19 of Graduate Studies in Mathematics. American Mathematical Society, Providence, RI, (2010), second edition. Google Scholar

• [51]

D. Gilbarg and N. S. Trudinger, Elliptic partial differential equations of second order, (2nd edition, revised 3rd printing). Springer Verlag, 1998. Google Scholar

• [52]

J. Malý and W. P. Ziemer, Fine regularity of solutions of elliptic partial differential equations, volume 51 of Mathematical Surveys and Monographs. American Mathematical Society, Providence, RI, 1997. Google Scholar

• [53]

L. Diening and F. Ettwein, Fractional estimates for non-differentiable elliptic systems with general growth. Forum Math., 20 (3), (2008), 523–556. Google Scholar

• [54]

P. Harjulehto, P. Hästö and A. Karppinen, Local higher integrability of the gradient of a quasiminimizer under generalized Orlicz growth conditions. Nonlinear Anal., 177, (2018), 543–552.

• [55]

E. Giusti, Direct methods in the calculus of variations. World Scientific Publishing Co. Inc., River Edge, NJ, 2003. Google Scholar

• [56]

L. Diening, P. Harjulehto, P. Hästö and M. Ružička, Lebesgue and Sobolev spaces with variable exponents. Lecture Notes in Mathematics, 2017, (2011), Springer-Verlag, Heidelberg. Google Scholar

• [57]

M. Giaquinta and G. Modica, Regularity results for some classes of higher order nonlinear elliptic systems. J. Reine Angew. Math., 311/312, (1979), 145–169. Google Scholar

• [58]

C. De Filippis and J. Oh, Regularity for multi-phase variational problems. Submitted. https://arxiv.org/abs/1807.02880Google Scholar

• [59]

C. De Filippis and G. Mingione, Manifold constrained non-uniformly elliptic problems. PreprintGoogle Scholar

• [60]

M. Giaquinta and L. Martinazzi, An introduction to the regularity theory for elliptic systems, harmonic maps and minimal graphs, volume 2 of Appunti. Scuola Normale Superiore di Pisa (Nuova Serie) [Lecture Notes. Scuola Normale Superiore di Pisa (New Series)]. Edizioni della Normale, Pisa, 2005. Google Scholar

Accepted: 2019-03-02

Published Online: 2019-07-20

Published in Print: 2019-03-01

Citation Information: Advances in Nonlinear Analysis, Volume 9, Issue 1, Pages 710–728, ISSN (Online) 2191-950X, ISSN (Print) 2191-9496,

Export Citation