EXISTENCE RESULT FOR STRONGLY NONLINEAR ELLIPTIC EQUATIONS IN ORLICZ SOBOLEV SPACES


 Archibald Hall
 1 years ago
 Views:
Transcription
1 Proyecciones Vol. 26, N o 2, pp , August Universidad Católica del Norte Antofagasta  Chile EXISTENCE RESULT FOR STRONGLY NONLINEAR ELLIPTIC EQUATIONS IN ORLIC SOBOLEV SPACES A. YOUSSFI UNIVERSITÉ SIDI MOHAMMED BEN ABDALLAH, MAROC Received : July Accepted : May 2007 Abstract In this paper, we prove the existence of solutions for some strongly nonlinear Dirichlet problems whose model is the following div(m 1 M( u ) u u )+um( u ) =f divf in D0 (), where is an open bounded subset of IR N,N 2. We emphasize that no 2 condition is required for the Nfunction M. Mathematics Subject Classification : (2000): 46E30, 35J60, 35J65. Key words: OrliczSobolev spaces, strongly nonlinear problems.
2 158 A. Youssfi 1. Introduction Let be a bounded open set of IR N, N 2, and let M be an Nfunction. Consider the following Dirichlet problem (1.1) A(u)+H(x, u, u) =f, where A(u) := div a(x, u, u) is a LerayLions type operator definedonitsdomaind(a) W 1 0 L M() and H is a nonlinearity assumed to satisfy the natural growth condition (1.2) H(x, s, ξ) b( s )(h(x)+m( ξ )) Recently, a large number of papers was devoted to the existence of solutions of (1.1). In the variational framework, that is f W 1 E M (), an existence result was proved in [8] when H depends only on x and u and satisfy the following sign condition H(x, s)s 0, andin[2]whenm satisfies the 2 condition and H depends also on u and satisfies (1.3) H(x, s, ξ)s 0. The result in [2] was generalized in [7] to Nfunctions without 2  condition. In the case where f L 1 (), problem (1.1) was solved in [3] under the socalled coercivity condition (1.4) H(x, s, ξ) βm( ξ ) for s some τ and in [5] assuming the sign condition (1.3) but the result was restricted to Nfunctions satisfying the 2 condition (see bellow). The result contained in [5] was then extended in [6] to Nfunctions without assuming the 2  condition. The solution u given in this case is such that its truncated
3 Existence result for strongly nonlinear elliptic equations in function T k (u) belongs to the energy space W 1 0 L M() for all k>0, but not the function u it self. Our main goal in this paper, is to prove the existence of a solution in W 1 0 L M() for problems of the kind of (1.1) when the source term has the form f divf with f L 1 () and F E M (), without any restriction on the Nfunction M. The paper is organized as follows, after giving a background in section 2, in section 3 we list the basic assumptions and our main result which will be proved in six steeps in section Prerequisites 2.1 Let M : IR + IR + be an Nfunction, ie. M is continuous, convex, with M(t) > 0fort>0, M(t) t 0ast 0and M(t) t as t.the Nfunction conjugate to M is defined as M(t) = sup{st M(t),s 0}. We recall the Young s inequality: for all s, t 0, If for some k>0, st M(s)+M(t). (2.1) M(2t) km(t) for all t 0, we said that M satisfies the 2 condition, and if (2.1) holds only for t some t 0, then M is said to satisfy the 2 condition near infinity. We will extend these Nfunctions into even functions on all IR. Let P and Q be two Nfunctions. the notation P Q means that P grows essentially less rapidly than Q, i.e. for all >0, that is the case if and only if P (t) Q( t) 0 as t, Q 1 (t) P 1 (t) 0 as t. 2.2 Let be an open subset of IR N. The Orlicz class K M () (resp. the Orlicz space L M ()) is defined as the set of (equivalence class of) realvalued measurable functions u on such that:
4 160 A. Youssfi µ u(x) M(u(x))dx < (resp. M dx < for some λ>0). λ Endowed with the Luxemburg norm kuk M =inf{λ >0: µ u(x) M dx < }, λ L M () is a Banach space and K M () is a convex subset of L M (). The Orlicz norm is defined on L M () by kuk (M) =sup u(x)v(x)dx, where the supremum is taken over all functions v L M () such that kvk M 1. The two norms k.k M and k.k (M) are equivalent (see [13]). The closure in L M () of the set of bounded measurable functions with compact support in is denoted by E M (). 2.3 The OrliczSobolev space W 1 L M () (resp.w 1 E M ()) is the space of functions u such that u and its distributional derivatives up to order 1 lie in L M () (resp.e M ()). It is a Banach space under the norm kuk 1,M = X kd α uk M. α 1 Thus, W 1 L M () andw 1 E M () canbeidentified with subspaces of the product of (N +1) copiesof L M (). Denoting this product by ΠL M,we will use the weak topologies σ(πl M, ΠE M )andσ(πl M, ΠL M ). The space W 1 0 E M() isdefined as the norm closure of the Schwartz space D() inw 1 E M () and the space W 1 0 L M() astheσ(πl M, ΠE M )closure of D() inw 1 L M (). We say that a sequence {u n } converges to u for the modular convergence in W 1 L M () if,forsomeλ>0, µ D α u n D α u M dx 0 for all α 1, λ this implies convergence for σ(πl M, ΠL M )(see[9,lemma6]). If M satisfies the 2 condition on IR + (near infinity only if has finite measure), then the modular convergence coincides with norm convergence (see [13, Theorem 9.4]).
5 Existence result for strongly nonlinear elliptic equations in Recall that the norm kduk M defined on W 1 0 L M() isequivalenttokuk 1,M (see [10]). Let W 1 L M () (resp. W 1 E M ()) denotes the space of distributions on which can be written as sums of derivatives of order 1 of functions in L M ()(resp. E M ()). It is a Banach space under the usual quotient norm. If the open has the segment property then the space D() isdense in W 1 0 L M() for the topology σ(πl M, ΠL M ) (see [10]). Consequently, the action of a distribution in W 1 L M () onanelementofw 1 0 L M() iswell defined. For an exhaustive treatments one can see for example [1, 13]. 2.4 We will use the following lemma, (see [6]), which concerns operators of Nemytskii Type in Orlicz spaces. It is slightly different from the analogous one given in [13]. Lemma 2.1. Let be an open subset of IR N with finite measure. let M, P and Q be Nfunctions such that Q P,andletf : IR IR be a Carathéodory function such that, for a.e.x and for all s IR, f(x, s) c(x)+k 1 P 1 M(k 2 s ), where k 1,k 2 are real constants and c(x) E Q (). Then the Nemytskii operator N f,defined by N f (u)(x) =f(x, u(x)), is strongly continuous from P(E M, k 1 2 )={u L M () :d(u, E M ()) < k 1 2 } into E Q (). We will use the following lemma which can be found in [12], Lemma 2.2. If {f n } L 1 () with f n f L 1 () a.e. in, f n,f 0 a.e. in and f n (x)dx f(x)dx, then f n f strongly in L 1 (). We also use the technical lemma: Lemma 2.3. Let x and y be two nonnegative real numbers and let with θ = y2 4x 2.Then φ(s) =se θs2, xφ 0 (s) y φ(s) x 2, s IR.
6 162 A. Youssfi 3. Assumptions and main result Let be an open bounded subset of IR N, N 2, with the segment property and let M and P be two Nfunctions such that P M. Let A : D(A) W0 1L M() W 1 L M () be a mapping (non everywhere defined) given by A(u) := div a(x, u, u) where a : IR IR N IR N is a Carathéodory function (i.e., a(x,, ) is continuous on IR IR N for almost every x in and a(,s,ξ)ismeasurable on for every (s, ξ) inir IR N ) satisfying for a.e. x, andforalls IR and all ξ, η IR N, ξ 6= η, (3.1) a(x, s, ξ) a 0 (x)+k 1 P 1 M(k 2 s )+k 1 M 1 M(k 2 ξ ) where a 0 (x) belongs to E M () andk 1,k 2 to IR +, (3.2) (a(x, s, ξ) a(x, s, η)) (ξ η) > 0 (3.3) a(x, s, ξ) ξ M( ξ ) Furthermore, let H : IR IR N IR be a Carathéodory function such that (3.4) H(x, s, ξ) b( s )(M( ξ )+h(x)) for almost x and for all s IR, ξ IR N,withb a real valued positive increasing continuous function and h a nonnegative function in L 1 (), and (3.5) H(x, s, ξ)sgn(s) M( ξ ) for a.e. x, foreveryξ IR N and for every s IR such that s σ, where σ is a positive real number. Consider the following Dirichlet problem: (3.6) A(u)+H(x, u, u) =f div(f ) in, u =0 on, We shall prove the following existence result:
7 Existence result for strongly nonlinear elliptic equations in Theorem 3.1. Assume that f L 1 (), F E M () and (3.1)(3.5) hold true, then there exists at least a function u solution of (3.6) in the sense that u W0 1L M(), H(x, u, u) L 1 () and a(x, u, u) T k (u v)dx + H(x, u, u)t k (u v)dx = ft k (u v)dx + F T k (u v)dx for every v W 1 0 L M() L () and every k σ. Remark We can replace assumptions (3.3), (3.4) and (3.5) by the following ones: (3.3) 0 a(x, s, ξ) ξ αm with α, λ > 0 and µ (3.4) 0 H(x, s, ξ) b( s ) M with 0 <λ μ and (3.5) 0 H(x, s, ξ)sgn(s) βm with 0 <τ λ and β>0. µ ξ λ µ ξ + h(x) μ µ ξ τ 2. A consequence of (3.3) and the continuity of a with respect to ξ, isthat, for almost every x in and s in IR, a(x, s, 0) = Note that assumption (3.5) gives a sign condition on H only near infinity. 4. In (3.4) we can assume only that b is positive and continuous. Remark 3.2. The solution of (3.6) given by theorem 3.1 belongs to W 1 0 L M() even if F =0, this regularity is due to assumption (3.5).
8 164 A. Youssfi 4. Proofoftheorem3.1 Let {f n } be a sequence of L () functions that converges strongly to f in L 1 (). Let n in IN and let H n (x, s, ξ) = H(x, s, ξ) 1+ 1 n H(x, s, ξ). It s easy to see that H n (x, s, ξ) n, H n (x, s, ξ) H(x, s, ξ) and H n (x, s, ξ)sgn(s) 0for s σ. Since H n is bounded for fixed n, there exists, (see [11, Propositions 1 and 5]), a function u n in W0 1L M() solution of A(u n )+H n (x, u n, u n )=f n divf in, u n =0 on, in the sense (3.7) a(x, u n, u n ) vdx + H n (x, u n, u n )vdx = f n vdx + F vdx for every v W 1 0 L M(). Step1: Estimation in W0 1L M(). For k>0, we denote by T k the usual truncation at level k defined by T k (s) =max( k, min(k, s)) for all s IR. Let us choose v = φ(t σ (u n )) as test function in (3.7), where σ is given by (3.5), φ is the function in lemma 2.3 and b is the function in (3.4). Using (3.3) and the Young s inequality, we obtain
9 Existence result for strongly nonlinear elliptic equations in M( T σ (u n ) )φ 0 (T σ (u n ))dx + H n (x, u n, u n )φ(t σ (u n ))dx φ(σ)kf n k L 1 () + φ 0 (σ) M(2 F )dx + 1 M( T σ (u n ) )φ 0 (T σ (u n ))dx. 2 Since {f n } is bounded in L 1 (), there exists a constant c not depending on n such that 1 M( T σ (u n ) )φ 0 (T σ (u n ))dx + H n (x, u n, u n )φ(t σ (u n ))dx 2 c(φ(σ)+φ 0 (σ)), which we can write, since H n enjoys the same properties of H, 1 M( T σ (u n ) )φ 0 (T σ (u n ))dx + H n (x, u n, u n )φ(t σ (u n ))dx 2 { u n <σ} + H n (x, u n, u n )φ(t σ (u n ))dx { u n σ} c(φ(σ)+φ 0 (σ)). By (3.4) we have H n (x, u n, u n )φ(t σ (u n ))dx { u n <σ} µ b(σ) M( T σ (u n ) )φ(t σ (u n ))dx + φ(σ)khk L 1 (), while using(3.5), we get { u n σ} H n (x, u n, u n )φ(t σ (u n ))dx φ(σ) M( u n )dx. { u n σ} Hence, we obtain R ³ M( T σ(u n ) ) 1 2 φ0 (T σ (u n )) b(σ) φ(t σ (u n )) dx +,φ(σ) M( u n )dx c(φ(σ)+φ 0 (σ)) + b(σ)φ(σ)khk L 1 (). { u n σ}
10 166 A. Youssfi Then, lemma 2.3 with the choice x = 1 2 and y = b(σ), yields 1 M( T σ (u n ) )dx + φ(σ) M( u n )dx 4 { u n σ} c(φ(σ)+φ 0 (σ)) + b(σ)φ(σ)khk L 1 (), which implies that (3.8) M( u n )dx c 0, where c 0 is a constant not depending on n. Thus {u n } is bounded in W0 1L M(), and consequently there exist a function u in W0 1L M() anda subsequence still denoted by {u n } such that (3.9) u n u in W 1 0 L M () for σ(πl M, ΠE M ) and (3.10) u n u in E M () strongly and a.e. in. Step2: {a(x, T k (u n ), T k (u n ))} is bounded in (L M ()) N for all k σ. We will use the Orlicz norm. For that, let ψ (L M ()) N with kψk M 1. For all k σ, we write using (3.2) so that µ a(x, T k (u n ), T k (u n )) a(x, T k (u n ), ψ ) µ T k (u n ) ψ dx 0, k 2 k2 (3.11) 1 a(x, T k (u n ), T k (u n )) ψdx k 2 a(x, T k (u n ), T k (u n )) T k (u n )dx + 1 k 2 a(x, T k (u n ), ψ k 2 ) T k (u n )dx a(x, T k (u n ), ψ k 2 ) ψdx.
11 Existence result for strongly nonlinear elliptic equations in To estimate the first term in the right, we take v = T k (u n )astest function in (3.7) and then use the Young s inequality, the fact that H n (x, u n, u n )T k (u n ) 0ontheset{ u n k} and (3.3), to obtain 1 a(x, T k (u n ), T k (u n )) T k (u n )dx + H n (x, u n, u n )T k (u n )dx 2 { u n k} kkf n k L 1 () + M(2 F )dx. Assumption (3.4) yields H n (x, u n, u n )T k (u n )dx { u n k} µ kb(k) M( u n )dx + khk L 1 () kb(k)(c 0 + khk L 1 ()), where c 0 is the constant in (3.8). Hence, since {f n } is bounded in L 1 (), we deduce that a(x, T k (u n ), T k (u n )) T k (u n )dx λ k, with λ k a constant depending on k. By the Young s inequality, (3.11) becomes 1 k 2 a(x, T k (u n ), T k (u n )) ψdx λ k +(1+2k 1 ) M( T k (u n ) )dx +(1+ 1 )(1 + 2k 1 ) M k k 1 k 2 M( ψ )dx. a(x, T k (u n ), ψ ) k 2 1+2k dx 1
12 168 A. Youssfi By virtue of (3.1) and the convexity of M, weget a(x, T k (u n ), ψ ) M k 2 1+2k dx 1 µ 1 ( a 0 (x) )dx + k 1 M 1 M(kk 2 ) 1+2k 1 We conclude that + k 1 M( ψ )dx. 1+2k 1 a(x, T k (u n ), T k (u n )) ψdx c k for all ψ (L M ()) N with kψk M 1, this means that (3.12) ka(x, T k (u n ), T k (u n ))k (M) c k, for every k σ. Step3: Almost everywhere convergence of the gradients. Since the function u belongs to W0 1L M(), there exists a sequence {v j } D(), (see [9]), which converges to u for the modular convergence in W0 1L M() anda.e.in. For m k σ, wedefine the function ρ m by 1 if s m ρ m (s) = m +1 s if m s m +1 0 if s m +1. Let θ j n = T k (u n ) T k (vj), θ j = T k (u) T k (v j )andz j n,m = φ(θ j n)ρ m (u n ) where φ is the function in lemma 2.3. In what follows, we denote by i (n, j), (i IN), various sequences of real numbers which tend to 0 when n and j respectively, i.e. lim j lim n i(n, j) =0.
13 Existence result for strongly nonlinear elliptic equations in We use z j n,m W 1 0 L M() as test function in (3.7) to get (3.13) <A(u n ),zn,m j > + H n (x, u n, u n )zn,mdx j = f n zn,mdx+ j F zn,mdx. j In view of (3.10), we have zn,m j φ(θ j )ρ m (u) weaklyinl () for σ (L,L 1 )asn,then lim f n z j n n,mdx = fφ(θ j )ρ m (u)dx, and since φ(θ j ) 0weaklyinL () forσ(l,l 1 )asj, we have lim fφ(θ j )ρ m (u)dx =0, j hence, we obtain f n zn,mdx j = 0 (n, j). Thanks to (3.8) and (3.10), we have as n z j n,m φ(θ j )ρ m (u) in W 1 0 L M () for σ(πl M, ΠE M ), which implies that lim F z j n n,mdx = F θ j φ 0 (θ j )ρ m (u)dx + F uφ(θ j )ρ 0 m(u)dx. On the one hand, by Lebesgue s theorem we get lim F uφ(θ j )ρ 0 m(u)dx =0, j on the other hand, we write F θ j φ 0 (θ j )ρ m (u)dx = F T k (u)φ 0 (θ j )ρ m (u)dx F T k (v j )φ 0 (θ j )ρ m (u)dx,
14 170 A. Youssfi so that, by Lebesgue s theorem one has lim F T k (u)φ 0 (θ j )ρ m (u)dx = F T k (u)ρ m (u)dx, j M ³ vj u Let λ>0 such that M λ 0 strongly in L 1 () asj and L 1 (), the convexity of the Nfunction M allows us to have ³ u λ M Ã Tk (v j )φ 0 (θ j! )ρ m (u) T k (u)ρ m (u) 4λφ 0 (2k) 1 µ 4 M vj u + 1 µ 1+ 1 λ 4 φ 0 M (2k) µ u λ. Then, by using the modular convergence of { v j } in (L M ()) N and Vitali s theorem, we obtain T k (v j )φ 0 (θ j )ρ m (u) T k (u)ρ m (u) in (L M ()) N for the modular convergence, and then lim F T k (v j )φ 0 (θ j )ρ m (u)dx = F T k (u)ρ m (u)dx. j We have proved that F z j n,mdx = 1 (n, j). Since H n (x, u n, u n )z j n,m 0ontheset{ u n >k} and ρ m (u n )=1on the set { u n k}, we have (3.14) <A(u n ),zn,m j > + H n (x, u n, u n )φ(θn)dx j 2 (n, j) { u n k} Now, we will evaluate the first term of the lefthand side of (3.14) by writing
15 Existence result for strongly nonlinear elliptic equations in <A(u n ),zn,m j > = a(x, u n, u n ) ( T k (u n ) T k (v j ))φ 0 (θn)ρ j m (u n )dx + = + and then a(x, u n, u n ) u n φ(θ j n)ρ 0 m(u n )dx a(x, T k (u n ), T k (u n )) ( T k (u n ) T k (v j ))φ 0 (θ j n)dx { u n >k} a(x, u n, u n ) T k (v j )φ 0 (θ j n)ρ m (u n )dx a(x, u n, u n ) u n φ(θ j n)ρ 0 m(u n )dx, (3.15) <A(u n ),zn,m j > = (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (v j )χ s j)) ( T k (u n ) T k (v j )χ s j )φ0 (θn)dx j + a(x, T k (u n ), T k (v j )χ s j)) ( T k (u n ) T k (v j )χ s j)φ 0 (θn)dx j a(x, T k (u n ), T k (u n )) T k (v j )φ 0 (θn)dx j \ s j a(x, u n, u n ) T k (v j )φ 0 (θn)ρ j m (u n )dx { u n >k} + a(x, u n, u n ) u n φ(θn)ρ j 0 m(u n )dx, where by χ s j, s>0, we denote the characteristic function of the subset s j = {x : T k (v j ) s}. For fixed m and s, we will pass to the limit in n andtheninj in the second, third, fourth and five terms in the right side of (3.15). Starting with the second term, we have
16 172 A. Youssfi a(x, T k (u n ), T k (v j )χ s j)) ( T k (u n ) T k (v j )χ s j)φ 0 (θ j n)dx a(x, T k (u), T k (v j )χ s j)) ( T k (u) T k (v j )χ s j)φ 0 (θ j )dx as n, since by lemma 2.1 one has a(x, T k (u n ), T k (v j )χ s j))φ 0 (θ j n) a(x, T k (u), T k (v j )χ s j))φ 0 (θ j ) strongly in (E M ()) N as n,while T k (u n ) T k (u) weakly in (L M ()) N by (3.8). Let χ s denote the characteristic function of the subset s = {x : T k (u) s}. As T k (v j )χ s j T k(u)χ s strongly in (E M ()) N as j, one has as j. Then a(x, T k (u), T k (v j )χ s j)) ( T k (u) T k (v j )χ s j)φ 0 (θ j )dx 0 (3.16) a(x, T k (u n ), T k (v j )χ s j)) ( T k (u n ) T k (v j )χ s j)φ 0 (θn)dx j = 3 (n, j). For the third term of (3.15), by virtue of (3.12) there exist a subsequence still indexed again by n and a function l k in (L M ()) N with k σ such that a(x, T k (u n ), T k (u n )) l k weakly in (L M ()) N for σ(πl M, ΠE M ). Then, since T k (v j )χ \ s j (E M ()) N,weobtain a(x, T k (u n ), T k (u n )) T k (v j )φ 0 (θn)dx j l k T k (v j )φ 0 (θ j )dx \ s j \ s j
17 Existence result for strongly nonlinear elliptic equations in as n. The modular convergence of {v j } allows us to have l k T k (v j )φ 0 (θ j )dx l k T k (u)dx \ s \ s j as j. This, proves that a(x, T k (u n ), T k (u n )) T k (v j )φ 0 (θn)dx j = l k T k (u)dx+ 4 (n, j). \ s j \ s (3.17) As regards the fourth term, observe that ρ m (u n ) = 0 on the subset { u n m +1}, sowehave a(x, u n, u n ) T k (v j )φ 0 (θn)ρ j m (u n )dx = { u n >k} { u n >k} a(x, T m+1 (u n ), T m+1 (u n )) T k (v j )φ 0 (θ j n)ρ m (u n )dx. As above, we obtain a(x, T m+1 (u n ), T m+1 (u n )) T k (v j )φ 0 (θn)ρ j m (u n )dx = { u n >k} { u >k} l m+1 T k (u)ρ m (u)dx + 5 (n, j). Observing that T k (u) = 0 on the subset { u >k}, one has (3.18) a(x, u n, u n ) T k (v j )φ 0 (θn)ρ j m (u n )dx = 5 (n, j) { u n >k} For the last term of (3.15), we have a(x, u n, u n ) u n φ(θn)ρ j 0 m(u n )dx = {m u n m+1} φ(2k) {m u n m+1} a(x, u n, u n ) u n φ(θn)ρ j 0 m(u n )dx a(x, u n, u n ) u n dx.
18 174 A. Youssfi To estimate the last term of the previous inequality, we test by T 1 (u n T m (u n )) W0 1L M() in (3.7), to get a(x, u n, u n ) u n dx {m u n m+1} + H n (x, u n, u n )T 1 (u n T m (u n ))dx { u n m} = f n T 1 (u n T m (u n ))dx + F u n dx. {m u n m+1} Using the fact that H n (x, u n, u n )T 1 (u n T m (u n )) 0 on the subset { u n m} and the Young s inequality, we get 1 a(x, u n, u n ) u n dx 2 {m u n m+1} f n dx + M( F )dx. { u n m} {m u n m+1} It follows that a(x, u n, u n ) u n φ(θn)ρ j 0 m(u n )dx (3.19) Ã! 2φ(2k) f n dx + M( F )dx. { u n m} {m u n m+1} From (3.16), (3.17), (3.18) and (3.19) we obtain (3.20) <A(u n ),zn,m j > (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (v j )χ s j)) ( T k (u n ) T k (v j )χ s j )φ0 (θ j n)dx Ã! 2φ(2k) f n dx + M( F )dx { u n m} {m u n m+1} \ s l k T k (u)dx + 6 (n, j).
19 Existence result for strongly nonlinear elliptic equations in Now, we turn to second term of the left hand side of (3.14). We have H n (x, u n, u n )φ(θ j n)dx { u n k} = H n (x, T k (u n ), T k (u n ))φ(θ j n)dx { u n k} b(k) M( T k (u n ) ) φ(θn) dx j + b(k) h(x) φ(θn) dx j b(k) a(x, T k (u n ), T k (u n )) T k (u n ) φ(θn) dx j + 7 (n, j). Then, (3.21) { u n k} b(k) H n (x, u n, u n )φ(θn)dx j ³ a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (v j )χ s j) ³ T k (u n ) T k (v j )χ s j φ(θn) dx j + b(k) a(x, T k (u n ), T k (v j )χ s j) ( T k (u n ) T k (v j )χ s j) φ(θn) dx j + b(k) a(x, T k (u n ), T k (u n )) T k (v j )χ s j φ(θn) dx j + 7 (n, j). We proceed as above to get b(k) a(x, T k (u n ), T k (v j )χ s j)) ( T k (u n ) T k (v j )χ s j) φ(θn) dx j = 8 (n, j) and b(k) a(x, T k (u n ), T k (u n )) T k (v j )χ s j φ(θn) dx j = 9 (n, j).
20 176 A. Youssfi Hence, we have (3.22) H n (x, u n, u n )φ(θ j n)dx { u n k} ³ b(k) a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (v j )χ s j) ³ T k (u n ) T k (v j )χ s j φ(θn) dx j + 10 (n, j). Combining (3.14), (3.20) and (3.22), we get ³ a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (v j )χ s j) ³ φ T k (u n ) T k (v j )χj s 0 (θn) j b(k) φ(θn) j dx l k T k (u)dx \ s Ã! +2φ(2k) f n dx + M( F )dx + 11 (n, j). { u n m} {m u n m+1} (3.23) Then, lemma 2.3 with x =1andy = b(k), yields ³ a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (v j )χ s j) ³ T k (u n ) T k (v j )χ s j dx 2 l k T k (u)dx \ s Ã! +4φ(2k) f n dx + M( F )dx + 11 (n, j). { u n m} {m u n m+1}
21 Existence result for strongly nonlinear elliptic equations in On the other hand (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (u)χ s )) ( T k (u n ) T k (u)χ s ) dx = ³ a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (v j )χ s j) ³ T k (u n ) T k (v j )χ s j dx + a(x, T k (u n ), T k (u n )) ( T k (v j )χ s j T k (u)χ s )dx a(x, T k (u n ), T k (u)χ s ) ( T k (u n ) T k (u)χ s )dx + a(x, T k (u n ), T k (v j )χ s j) ( T k (u n ) T k (v j )χ s j)dx. We shall pass to the limit in n andtheninj in the last three terms of the right hand side of the above equality. By similar arguments as in (3.15) and (3.21), we obtain a(x, T k (u n ), T k (u n )) ( T k (v j )χ s j T k (u)χ s )dx = 12 (n, j) and and a(x, T k (u n ), T k (u)χ s ) ( T k (u n ) T k (u)χ s )dx = 13 (n, j) (3.24) so that a(x, T k (u n ), T k (v j )χ s j) ( T k (u n ) T k (v j )χ s j)dx = 14 (n, j),
22 178 A. Youssfi (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (u)χ s )) ( T k (u n ) T k (u)χ s )dx (3.25) = (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (v j )χ s j)) ( T k (u n ) T k (v j )χ s j )dx + 15 (n, j). Let r s, we use (3.2), (3.25) and (3.23) to get 0 (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (u))) r ( T k (u n ) T k (u))dx (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (u))) s ( T k (u n ) T k (u))dx = (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (u)χ s )) s ( T k (u n ) T k (u)χ s )dx (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (u)χ s )) ( T k (u n ) T k (u)χ s )dx
23 Existence result for strongly nonlinear elliptic equations in = (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (v j )χ s j)) ( T k (u n ) T k (v j )χ s j)dx + 15 (n, j) 2 l k T k (u)dx \ s Ã! +4φ(2k) f n dx + M( F )dx { u n m} {m u n m+1} + 16 (n, j), Which gives by passing to the limit sup over n and then over j 0 lim sup (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (u))) n r ( T k (u n ) T k (u))dx Ã! 2 l k T k (u)dx +4φ(2k) f dx + M( F )dx. \ s { u m} {m u m+1} Letting s and then m, taking into account that l k T k (u) L 1 (), f L 1 (), M( F ) L 1 (), \ s 0, { u m} 0, and {m u m +1} 0, one has (a(x, T k (u n ), T k (u n )) a(x, T k (u n ), T k (u))) ( T k (u n ) T k (u))dx 0 r as n. As in [4], we deduce that there exists a subsequence of {u n } still indexed again by n such that (3.26) u n u a.e. in. Thus, by (3.12) and (3.26) we have (3.27) a(x, T k (u n ), T k (u n )) a(x, T k (u), T k (u)) weakly in (L M ()) N for σ(πl M, ΠE M )andforallk σ.
24 180 A. Youssfi Step4: Modular convergence of the truncations. Going back to (3.23), we write a(x, T k (u n ), T k (u n )) T k (u n )dx a(x, T k (u n ), T k (u n )) T k (v j )χ s jdx + a(x, T k (u n ), T k (v j )χ s j) ( T k (u n ) T k (v j )χ s j)dx Ã! +4φ(2k) f n dx + M( F )dx { u n m} {m u n m+1} +2 a(x, T k (u), T k (u)) T k (u)dx + 11 (n, j). \ s and by (3.24) we get a(x, T k (u n ), T k (u n )) T k (u n )dx a(x, T k (u n ), T k (u n )) T k (v j )χ s jdx Ã! +4φ(2k) f n dx + M( F )dx { u n m} {m u n m+1} +2 a(x, T k (u), T k (u)) T k (u)dx + 17 (n, j). \ s We pass now to the limit sup over n in both sides of this inequality, to obtain lim sup a(x, T k (u n ), T k (u n )) T k (u n )dx n a(x, T k (u), T k (u)) T k (v j )χ s jdx Ã! +4φ(2k) f dx + M( F )dx { u m} {m u m+1} +2 a(x, T k (u), T k (u)) T k (u)dx + lim 17(n, j), \ s n
25 Existence result for strongly nonlinear elliptic equations in in which, we pass to the limit in j to get lim sup a(x, T k (u n ), T k (u n )) T k (u n )dx n a(x, T k (u), T k (u)) T k (u)χ s dx Ã! +4φ(2k) f dx + M( F )dx { u m} {m u m+1} +2 a(x, T k (u), T k (u)) T k (u)dx, \ s letting s and then m, one has lim sup a(x, T k (u n ), T k (u n )) T k (u n )dx a(x, T k (u), T k (u)) T k (u)dx n On the other hand, by Fatou s lemma, we have a(x, T k (u), T k (u)) T k (u)dx lim inf a(x, T k (u n ), T k (u n )) T k (u n )dx. n It follows that a(x, T k (u n ), T k (u n )) T k (u n )dx = lim n By lemma 2.2 we conclude that a(x, T k (u), T k (u)) T k (u)dx. (3.28) a(x, T k (u n ), T k (u n )) T k (u n ) a(x, T k (u), T k (u)) T k (u) strongly in L 1 (), k σ. The convexity of the Nfunction M and (3.3) allow us to have µ Tk (u n ) T k (u) M a(x, T k(u n ), T k (u n )) T k (u n )+ 1 2 a(x, T k(u), T k (u)) T k (u). Then, by (3.28) we get µ lim sup Tk (u n ) T k (u) M dx =0. E 0 n E 2
26 182 A. Youssfi So that, by Vitali s theorem one has T k (u n ) T k (u) in W 1 0 L M () for the modular convergence, for all k σ. Step5: Equiintegrability of the nonlinearities. As a consequence of (3.10) and (3.26), one has H n (x, u n, u n ) H(x, u, u) a.e. in. We shall prove that the sequence {H n (x, u n, u n )} is uniformly equiintegrable in. Let E be a measurable subset of, for all m σ, wehave H n (x, u n, u n ) dx = H n (x, u n, u n ) dx E + H n (x, u n, u n ) dx. E { u n >m} E { u n m} On the one hand, the use of T 1 (u n T m 1 (u n )) as test function in (3.7), the Young s inequality and (3.3) led to H n (x, u n, u n )T 1 (u n T m 1 (u n ))dx f n dx + M(2 F )dx. {m 1 u n m} { u n m 1} { u n m} Then, assumption (3.5) gives H n (x, u n, u n ) dx { u n m 1} f n dx+ M(2 F )dx. {m 1 u n m} For all >0, one can find an m = m( ) > 1 such that sup H n (x, u n, u n ) dx n { u n m} 2.
27 Existence result for strongly nonlinear elliptic equations in E { u n m} On the other hand, we use (3.3) and (3.4) to get H n (x, u n, u n ) dx H n (x, T m (u n ), T m (u n )) dx E µ b(m) E M( T m (u n ) )dx + h(x)dx E b(m) a(x, T m (u n ), T m (u n )) T m (u n )dx E +b(m) h(x)dx. We use the fact that from (3.28) the sequence {a(x, T m (u n ), T m (u n )) T m (u n )} is equiintegrable and that h L 1 () toobtain lim sup H n (x, u n, u n ) dx =0, E 0 n E { u n m} where E denotes the Lebesgue measure of the subset E. Consequently lim sup H n (x, u n, u n ) dx =0. E 0 n E This proves that the sequence {H n (x, u n, u n )} is uniformly equiintegrable in. By Vitali s theorem, we conclude that H(x, u, u) L 1 () and (3.29) H n (x, u n, u n ) H(x, u, u) strongly in L 1 (). E Step6: Passage to the limit. Let v W 1 0 L M() L (). By [9, Lemma 4], there exists a sequence {v j } D() such that kv j k (N +1)kvk and v j v in W 1 0 L M () for the modular convergence and a.e. in. Let k σ. We go back to approximate equations (3.7) and use T k (u n v j ) as test function to obtain (3.30) a(x, T t (u n ), T t (u n )) T k (u n v j )dx + H n (x, u n, u n )T k (u n v j )dx = f n T k (u n v j )dx + F T k (u n v j )dx,
28 184 A. Youssfi where t = k +(N +1)kvk. The first term in the lefthand side of (7.3.23) is written as a(x, T t (u n ), T t (u n )) T k (u n v j )dx = { u n v j <k} { u n v j <k} Thus, by (3.27) and (3.28) we obtain a(x, T t (u n ), T t (u n )) T k (u n v j )dx as n. Since a(x, T t (u n ), T t (u n )) T t (u n )dx a(x, T t (u n ), T t (u n )) v j dx a(x, T t (u), T t (u)) T k (u v j )dx T k (u n v j ) T k (u v j ) in L () forσ (L,L 1 ), we use (3.29) and the fact that f n f strongly in L 1 () asn, to obtain H n (x, u n, u n )T k (u n v j )dx H(x, u, u)t k (u v j )dx, f n T k (u n v j )dx ft k (u v j )dx, as n. For the last term in the righthand side of (3.30) we write F T k (u n v j )dx = F u n dx F v j dx. { u n v j <k} { u n v j <k} Hence, by (3.9) we obtain F T k (u n v j )dx F T k (u v j )dx. Therefore, passing to the limit as n in (3.30), we get a(x, u, u) T k (u v j )dx + H(x, u, u)t k (u v j )dx = ft k (u v j )dx + F T k (u v j )dx,
29 Existence result for strongly nonlinear elliptic equations in in which we pass to the limit as j to obtain a(x, u, u) T k (u v)dx + H(x, u, u)t k (u v)dx = ft k (u v)dx + F T k (u v)dx. Which completes the proof of theorem 3.1. Remark 4.1. If the Nfunction M satisfies the 2 condition, the sequence {a(x, u n, u n )} will be bounded in (L M ()) N. Then, the function u solution of the problem (3.6) is such that: u W0 1L M(), H(x, u, u) L 1 () and a(x, u, u) vdx + H(x, u, u)vdx = fvdx + F vdx, for every v W 1 0 L M() L (). Remark 4.2. We can interpret theorem 3.1 in the following sense: the problem ( u W 1 0 L M (), H(x, u, u) L 1 () div a(x, u, u)+h(x, u, u) =μ admits a solution if and only if μ belongs to L 1 ()+W 1 L M (). Remark 4.3. If we replace (3.1) by the more general growth condition a(x, s, ξ) b 0 ( s )(a 0 (x)+m 1 M(τ ξ )) where a 0 (x) belongs to E M (), τ > 0 and b 0 is a positive continuous increasing function, we can adapt the same ideas to prove the existence of solutions for the problem u W0 1L M(), H(x, u, u) L 1 () and a(x, u, u) T k (u v)dx + H(x, u, u)t k (u v)dx = ft k (u v)dx + F T k (u v)dx for v W0 1L M() L (), by considering the following approximation problems ( un W0 1L M() div a(x, T n (u n ), u n )+H n (x, u n, u n )=f n divf in.
30 186 A. Youssfi As an application of this result, we give div((1 + u ) q exp( u p ) 1 u 2 u)+u(exp( u p ) 1) = f divf in. with p>1 and q>0. References [1] Adams, R.: Sobolev spaces, Academic Press Inc, New York, (1975) [2] Benkirane, A., Elmahi, A.: An existence theorem for a strongly nonlinear problems in Orlicz spaces, Nonlinear Anal. T.M.A. 36, pp (1999). [3] Benkirane, A., Elmahi, A.: strongly nonlinear elliptic equations having natural growth terms and L 1 data, Nonlinear Anal. T.M.A. 39, pp (2000). [4] Benkirane, A., Elmahi, A.: Almost everywhere convergence of the gradients of solutions to elliptic equations in Orlicz spaces and application, Nonlinear Anal. T.M.A. 28, pp (1997). [5] Benkirane, A., Elmahi, A., Meskine, D.: An existence theorem for a class of elliptic problems in L 1, Applicationes Mathematicae 29, 4, pp (2002). [6] Elmahi, A., Meskine, D.: Nonlinear elliptic problems having natural growth and L 1 data in Orlicz spaces, Ann. Mat. Pura Appl. 184, 2, pp (2004). [7] Elmahi, A., Meskine, D.: Existence of solutions for elliptic equations having natural growth terms in Orlicz spaces, Abst. Appl. Anal. 2004, 12, pp (2004). [8] Gossez, J.P.: A strongly nonlinear elliptic problem in OrliczSobolev spaces, Proc. Sympos. Pure Math., 45, Part 1, (1986). [9] Gossez, J.P.: Some approximation properties in OrliczSobolev spaces, Stud. Math. 74, pp (1982).
A UNIQUENESS RESULT FOR THE CONTINUITY EQUATION IN TWO DIMENSIONS. Dedicated to Constantine Dafermos on the occasion of his 70 th birthday
A UNIQUENESS RESULT FOR THE CONTINUITY EQUATION IN TWO DIMENSIONS GIOVANNI ALBERTI, STEFANO BIANCHINI, AND GIANLUCA CRIPPA Dedicated to Constantine Dafermos on the occasion of his 7 th birthday Abstract.
More informationProperties of BMO functions whose reciprocals are also BMO
Properties of BMO functions whose reciprocals are also BMO R. L. Johnson and C. J. Neugebauer The main result says that a nonnegative BMOfunction w, whose reciprocal is also in BMO, belongs to p> A p,and
More informationAn existence result for a nonconvex variational problem via regularity
An existence result for a nonconvex variational problem via regularity Irene Fonseca, Nicola Fusco, Paolo Marcellini Abstract Local Lipschitz continuity of minimizers of certain integrals of the Calculus
More informationControllability and Observability of Partial Differential Equations: Some results and open problems
Controllability and Observability of Partial Differential Equations: Some results and open problems Enrique ZUAZUA Departamento de Matemáticas Universidad Autónoma 2849 Madrid. Spain. enrique.zuazua@uam.es
More informationA SURGERY RESULT FOR THE SPECTRUM OF THE DIRICHLET LAPLACIAN. Keywords: shape optimization, eigenvalues, Dirichlet Laplacian
A SURGERY RESULT FOR THE SPECTRUM OF THE DIRICHLET LAPLACIA DORI BUCUR AD DARIO MAZZOLEI Abstract. In this paper we give a method to geometrically modify an open set such that the first k eigenvalues of
More informationA new continuous dependence result for impulsive retarded functional differential equations
CADERNOS DE MATEMÁTICA 11, 37 47 May (2010) ARTIGO NÚMERO SMA#324 A new continuous dependence result for impulsive retarded functional differential equations M. Federson * Instituto de Ciências Matemáticas
More informationLaw Invariant Risk Measures have the Fatou Property
Law Invariant Risk Measures have the Fatou Property E. Jouini W. Schachermayer N. Touzi Abstract S. Kusuoka [K 1, Theorem 4] gave an interesting dual characterization of law invariant coherent risk measures,
More informationONEDIMENSIONAL RANDOM WALKS 1. SIMPLE RANDOM WALK
ONEDIMENSIONAL RANDOM WALKS 1. SIMPLE RANDOM WALK Definition 1. A random walk on the integers with step distribution F and initial state x is a sequence S n of random variables whose increments are independent,
More informationDecoding by Linear Programming
Decoding by Linear Programming Emmanuel Candes and Terence Tao Applied and Computational Mathematics, Caltech, Pasadena, CA 91125 Department of Mathematics, University of California, Los Angeles, CA 90095
More informationL (with coefficients oscillating with period e), and a family of solutions u which,
SIAM J. MATH. ANAL. Vol. 23, No. 6, pp. 14821518, November 1992 (C) 1992 Society for Industrial and Applied Mathematics 006 HOMOGENIZATION AND TWOSCALE CONVERGENCE* GRIGOIRE ALLAIRE Abstract. Following
More informationSecondorder conditions in C 1,1 constrained vector optimization
Ivan Ginchev, Angelo Guerreggio, Matteo Rocca Secondorder conditions in C 1,1 constrained vector optimization 2004/17 UNIVERSITÀ DELL'INSUBRIA FACOLTÀ DI ECONOMIA http://eco.uninsubria.it In questi quaderni
More informationEMPIRICAL PROCESSES: THEORY AND APPLICATIONS
NSFCBMS Regional Conference Series in Probability and Statistics Volume 2 EMPIRICAL PROCESSES: THEORY AND APPLICATIONS David Pollard Yale University Sponsored by the Conference Board of the Mathematical
More informationSubspace Pursuit for Compressive Sensing: Closing the Gap Between Performance and Complexity
Subspace Pursuit for Compressive Sensing: Closing the Gap Between Performance and Complexity Wei Dai and Olgica Milenkovic Department of Electrical and Computer Engineering University of Illinois at UrbanaChampaign
More informationA Modern Course on Curves and Surfaces. Richard S. Palais
A Modern Course on Curves and Surfaces Richard S. Palais Contents Lecture 1. Introduction 1 Lecture 2. What is Geometry 4 Lecture 3. Geometry of InnerProduct Spaces 7 Lecture 4. Linear Maps and the Euclidean
More informationOrthogonal Bases and the QR Algorithm
Orthogonal Bases and the QR Algorithm Orthogonal Bases by Peter J Olver University of Minnesota Throughout, we work in the Euclidean vector space V = R n, the space of column vectors with n real entries
More informationLAGUERRE POLYNOMIALS WITH GALOIS GROUP A m FOR EACH m
LAGUERRE POLYNOMIALS WITH GALOIS GROUP A m FOR EACH m Michael Filaseta Mathematics Department University of South Carolina Columbia, SC 29208 Ognian Trifonov Mathematics Department University of South
More informationNotes on Richard Dedekind s Was sind und was sollen die Zahlen?
Notes on Richard Dedekind s Was sind und was sollen die Zahlen? David E. Joyce, Clark University December 2005 Contents Introduction 2 I. Sets and their elements. 2 II. Functions on a set. 5 III. Onetoone
More informationConsistency and Convergence Rates of OneClass SVMs and Related Algorithms
Journal of Machine Learning Research 7 2006 817 854 Submitted 9/05; Revised 1/06; Published 5/06 Consistency and Convergence Rates of OneClass SVMs and Related Algorithms Régis Vert Laboratoire de Recherche
More informationAn Elementary Introduction to Modern Convex Geometry
Flavors of Geometry MSRI Publications Volume 3, 997 An Elementary Introduction to Modern Convex Geometry KEITH BALL Contents Preface Lecture. Basic Notions 2 Lecture 2. Spherical Sections of the Cube 8
More informationGeneralized compact knapsacks, cyclic lattices, and efficient oneway functions
Generalized compact knapsacks, cyclic lattices, and efficient oneway functions Daniele Micciancio University of California, San Diego 9500 Gilman Drive La Jolla, CA 920930404, USA daniele@cs.ucsd.edu
More informationRegular Languages are Testable with a Constant Number of Queries
Regular Languages are Testable with a Constant Number of Queries Noga Alon Michael Krivelevich Ilan Newman Mario Szegedy Abstract We continue the study of combinatorial property testing, initiated by Goldreich,
More informationErgodicity and Energy Distributions for Some Boundary Driven Integrable Hamiltonian Chains
Ergodicity and Energy Distributions for Some Boundary Driven Integrable Hamiltonian Chains Peter Balint 1, Kevin K. Lin 2, and LaiSang Young 3 Abstract. We consider systems of moving particles in 1dimensional
More informationAN OVERVIEW OF SOME STOCHASTIC STABILITY METHODS
Journal of the Operations Research Society of Japan 2004, Vol. 47, No. 4, 275303 AN OVERVIEW OF SOME STOCHASTIC STABILITY METHODS Serguei Foss Takis Konstantopoulos HeriottWatt University (Received December
More informationHow to Use Expert Advice
NICOLÒ CESABIANCHI Università di Milano, Milan, Italy YOAV FREUND AT&T Labs, Florham Park, New Jersey DAVID HAUSSLER AND DAVID P. HELMBOLD University of California, Santa Cruz, Santa Cruz, California
More informationIEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 52, NO. 4, APRIL 2006 1289. Compressed Sensing. David L. Donoho, Member, IEEE
IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 52, NO. 4, APRIL 2006 1289 Compressed Sensing David L. Donoho, Member, IEEE Abstract Suppose is an unknown vector in (a digital image or signal); we plan to
More informationDiscreteTime Signals and Systems
2 DiscreteTime Signals and Systems 2.0 INTRODUCTION The term signal is generally applied to something that conveys information. Signals may, for example, convey information about the state or behavior
More informationProbability in High Dimension
Ramon van Handel Probability in High Dimension ORF 570 Lecture Notes Princeton University This version: June 30, 2014 Preface These lecture notes were written for the course ORF 570: Probability in High
More informationAN INTRODUCTION TO OMINIMAL GEOMETRY. Michel COSTE Institut de Recherche Mathématique de Rennes michel.coste@univrennes1.fr
AN INTRODUCTION TO OMINIMAL GEOMETRY Michel COSTE Institut de Recherche Mathématique de Rennes michel.coste@univrennes1.fr November 1999 Preface These notes have served as a basis for a course in Pisa
More informationFirstorder definable languages
Firstorder definable languages Volker Diekert 1 Paul Gastin 2 1 Institut für Formale Methoden der Informatik Universität Stuttgart Universitätsstraße 38 70569 Stuttgart, Germany diekert@fmi.unistuttgart.de
More informationSpaceTime Approach to NonRelativistic Quantum Mechanics
R. P. Feynman, Rev. of Mod. Phys., 20, 367 1948 SpaceTime Approach to NonRelativistic Quantum Mechanics R.P. Feynman Cornell University, Ithaca, New York Reprinted in Quantum Electrodynamics, edited
More information