Electronic Journal of Differential Equations, Vol. 2012 (2012), No. 163, pp. 1–9. ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu ftp ejde.math.txstate.edu INFINITELY MANY SOLUTIONS FOR CLASS OF NAVIER BOUNDARY (p, q)-BIHARMONIC SYSTEMS MOHAMMED MASSAR, EL MILOUD HSSINI, NAJIB TSOULI Abstract. This article shows the existence and multiplicity of weak solutions for the (p, q)-biharmonic type system ∆(|∆u|p−2 ∆u) = λFu (x, u, v) ∆(|∆v| q−2 ∆v) = λFv (x, u, v) u = v = ∆u = ∆v = 0 in Ω, in Ω, on ∂Ω. Under certain conditions on F , we show the existence of infinitely many weak solutions. Our technical approach is based on Bonanno and Molica Bisci’s general critical point theorem. 1. Introduction In this paper we are concerned with the existence and multiplicity of weak solutions for the (p, q)-biharmonic type system ∆(|∆u|p−2 ∆u) = λFu (x, u, v) q−2 ∆(|∆v| ∆v) = λFv (x, u, v) u = v = ∆u = ∆v = 0 in Ω, in Ω, (1.1) on ∂Ω, N where Ω is an open bounded subset of R (N ≥ 1), with smooth boundary, λ ∈ (0, ∞), p > max{1, N2 }, q > max{1, N2 }. F : Ω × R2 → R is a function such that F (., s, t) is continuous in Ω, for all (s, t) ∈ R2 and F (x, ., .) is C 1 in R2 for every x ∈ Ω, and Fu , Fv denote the partial derivatives of F , with respect to u, v respectively. The investigation of existence and multiplicity of solutions for problems involving biharmonic and p-biharmonic operators has drawn the attention of many authors, see [5, 9, 12, 15] and references therein. Candito and Livrea [5] considered the nonlinear elliptic Navier boundary-value problem ∆(|∆u|p−2 ∆u) = λf (x, u) u = ∆u = 0 i nΩ, on ∂Ω. There the authors established the existence of infinitely many solutions. 2000 Mathematics Subject Classification. 35J40, 35J60. Key words and phrases. Navier value problem; infinitely many solutions; Ricceri’s variational principle. c 2012 Texas State University - San Marcos. Submitted June 4, 2012. Published September 21, 2012. 1 (1.2) 2 M. MASSAR, E. M. HSSINI, N. TSOULI EJDE-2012/163 In the present paper, we look for the existence of infinitely many solutions of system (1.1). More precisely, we will prove the existence of well precise intervals of parameters such that problem (1.1) admits either an unbounded sequence of solutions provided that F (x, u, v) has a suitable behaviour at infinity or a sequence of nontrivial solutions converging to zero if a similar behaviour occurs at zero. Our main tool is a general critical points theorem due to Bonanno and Molica Bisci [2] that is a generalization of a previous result of Ricceri [11]. In the sequel, X will denote the space W 2,p (Ω) ∩ W01,p (Ω) × W 2,q (Ω) ∩ W01,q (Ω) , which is a reflexive Banach space endowed with the norm k(u, v)k = kukp + kvkq , where kukp = Z p |∆u| dx 1/p and kvkq = Z Ω |∆v|q dx 1/q . Ω Let K := max n max|u(x)|p x∈Ω sup u∈W 2,p (Ω)∩W01,p (Ω)\{0} , kukpp max|v(x)|q o x∈Ω . 1,q kvkqq 2,q v∈W (Ω)∩W (Ω)\{0} sup 0 (1.3) Since p > max{1, N2 } and q > max{1, N2 }, the Rellich Kondrachov theorem assures that W 2,p (Ω) ∩ W01,p (Ω) ,→ C(Ω) and W 2,q (Ω) ∩ W01,q (Ω) ,→ C(Ω) are compact, and hence K < ∞. Definition 1.1. We say that (u, v) ∈ X is a weak solution of problem (1.1) if Z Z p−2 |∆u| ∆u∆ϕ dx + |∆v|q−2 ∆v∆ψ dx Ω Ω Z Z −λ Fu (x, u, v)ϕ dx − λ Fv (x, u, v)ψ dx = 0, Ω Ω for all (ϕ, ψ) ∈ X. Define the functional Iλ : X → R, given by Iλ (u, v) = Φ(u, v) − λΨ(u, v), for all (u, v) ∈ X, where 1 1 Φ(u, v) = kukpp + kvkqq p q Z and Ψ(u, v) = F (x, u, v)dx. Ω Since X is compactly embedded in C 0 (Ω) × C 0 (Ω), it is well known that Φ and Ψ are well defined Gâteaux differentiable functionals whose Gâteaux derivatives at (u, v) ∈ X are given by Z Z hΦ0 (u, v), (ϕ, ψ)i = |∆u|p−2 ∆u∆ϕdx + |∆v|q−2 ∆v∆ψdx, Ω Z ZΩ 0 hΨ (u, v), (ϕ, ψ)i = Fu (x, u, v)ϕdx + Fv (x, u, v)ψdx, Ω Ω for all (ϕ, ψ) ∈ X. Moreover, by the weakly lower semicontinuity of norm, we see that Φ is sequentially weakly lower semi continuous. Since Ψ has compact derivative, it follows that Ψ is sequentially weakly continuous. EJDE-2012/163 INFINITELY MANY SOLUTIONS 3 In view of (1.3), for every (u, v) ∈ X, we have sup |u(x)|p ≤ Kkukpp and x∈Ω thus sup |v(x)|q ≤ Kkvkqq , x∈Ω 1 1 1 |u(x)|p + |v(x)|q ≤ K kukpp + kvkqq . (1.4) q p q x∈Ω p Hence, for every r > 0 Φ−1 (] − ∞, r[) : = (u, v) ∈ X : Φ(u, v) < r 1 1 = (u, v) ∈ X : kukpp + kvkqq < r (1.5) p q 1 1 ⊆ (u, v) ∈ X : |u(x)|p + |v(x)|q < Kr, ∀ x ∈ Ω . p q Let us recall for the reader’s convenience a smooth version of a previous result of Ricceri [11]. sup 1 Theorem 1.2. Let X be a reflexive real Banach space, let Φ, Ψ : X → R be two Gâteaux differentiable functionals such that Φ is sequentially weakly lower semicontinuous and coercive and Ψ is sequentially weakly upper semicontinuous. For every r > inf X Φ, let us put supv∈Φ−1 (]−∞,r[) Ψ(v) − Ψ(u) ϕ(r) := inf r − Φ(u) u∈Φ−1 (]−∞,r[) and γ := lim inf ϕ(r), r→+∞ δ := lim inf r→(inf X Φ)+ ϕ(r). Then, one has 1 (a) for every r > inf X Φ and every λ ∈]0, ϕ(r) [, the restriction of the functional −1 Iλ = Φ − λΨ to Φ (] − ∞, r[) admits a global minimum, which is a critical point (local minimum) of Iλ in X. (b) If γ < +∞ then, for each λ ∈]0, γ1 [, the following alternative holds: either (b1) Iλ possesses a global minimum,or (b2) there is a sequence (un ) of critical points (local minima) of Iλ such that limn→+∞ Φ(un ) = +∞. (c) If δ < +∞ then, for each λ ∈]0, 1δ [, the following alternative holds: either (c1) there is a global minimum of Φ which is a local minimum of Iλ ,or (c2) there is a sequence of pairwise distinct critical points (local minima) of Iλ which weakly converges to global minimum of Φ. 2. Main results Fix x0 ∈ Ω and pick R2 > R1 > 0 such that B(x0 , R2 ) ⊆ Ω. Set R 2 − R 2 p 1 Γ(1 + N/2) 2 1 , Lp := min(p,q) N − RN 2N R 1/p 1/q N/2 2 1 (Kp) + (Kq) π R 2 − R 2 q Γ(1 + N/2) 1 2 1 Lq := min(p,q) N − RN 2N R 1/p 1/q N/2 2 1 (Kp) + (Kq) π (2.1) where Γ denotes the Gamma function and K is given by (1.3). Now we are ready to state our main results. 4 M. MASSAR, E. M. HSSINI, N. TSOULI EJDE-2012/163 Theorem 2.1. Assume that (i1) F (x, s, t) ≥ 0 for every (x, s, t) ∈ Ω × [0, +∞)2 ; (i2) There exist x0 ∈ Ω, 0 < R1 < R2 as considered in (2.1) such that, if we put R R F (x, s, t)dx sup|s|+|t|≤σ F (x, s, t)dx B(x0 ,R1 ) Ω , β := lim sup α := lim inf , p s tq min(p,q) σ→+∞ σ s,t→+∞ p + q one has α < Lβ, (2.2) where L := min{Lp , Lq }. Then, for every λ ∈ Λ := 1 1 min(p,q) Lβ , α (Kp)1/p + (Kq)1/q 1 problem (1.1) admits an unbounded sequence of weak solutions. Theorem 2.2. Assume that (i1) holds and (i3) F (x, 0, 0) = 0 for every x ∈ Ω. (i4) There exist x0 ∈ Ω, 0 < R1 < R2 as considered in (2.1) such that, if we put R R F (x, s, t)dx sup F (x, s, t)dx |s|+|t|≤σ B(x0 ,R1 ) 0 , α0 := lim inf Ω , β := lim sup p s tq σ→0+ σ min(p,q) s,t→0+ p + q one has α0 < Lβ 0 . (2.3) where L := min{Lp , Lq }. Then, for every λ ∈ Λ := 1 1 1 min(p,q) Lβ 0 , α0 , (Kp)1/p + (Kq)1/q problem (1.1) admits a sequence (un ) of weak solutions such that un * 0. 3. Proofs of main results Proof of Theorem 2.1. To apply Theorem 1.2, we set sup(w,z)∈Φ−1 (]−∞,r[) Ψ(w, z) − Ψ(u, v) ϕ(r) := inf r − Φ(u, v) (u,v)∈Φ−1 (]−∞,r[) Note that Φ(0, 0) = 0, and by (i1), Ψ(0, 0) ≥ 0. Therefore, for every r > 0, sup(w,z)∈Φ−1 (]−∞,r[) Ψ(w, z) − Ψ(u, v) ϕ(r) = inf r − Φ(u, v) (u,v)∈Φ−1 (]−∞,r[) supΦ−1 (]−∞,r[) Ψ ≤ r R supΦ(u,v)<r Ω F (x, u, v)dx = . r (3.1) EJDE-2012/163 INFINITELY MANY SOLUTIONS 5 Hence, from (1.5), we have 1 ϕ(r) ≤ sup r {(u,v)∈X: |u(x)|p + |v(x)|q <Kr, ∀ x∈Ω} p Z F (x, u, v)dx Ω q Let (σn ) a sequence of positive numbers such that σn → +∞ and R sup|s|+|t|≤σn F (x, s, t)dx lim Ω = α < +∞. min(p,q) n→+∞ σn (3.2) Put rn := min(p,q) σn (Kp)1/p + (Kq)1/q Let (u, v) ∈ Φ−1 (] − ∞, rn [), from (1.5) we have |v(x)|q |u(x)|p + < Krn , p q for all x ∈ Ω. Thus |u(x)| ≤ (Kprn )1/p and |v(x)| ≤ (Kqrn )1/q , hence, for n large enough (rn > 1), |u(x)| + |v(x)| ≤ (Kprn )1/p + (Kqrn )1/q 1 ≤ (Kp)1/p + (Kq)1/q rnmin(p,q) = σn . Therefore, R sup{(u,v)∈X:|u(x)|+|v(x)|<σn , ∀ x∈Ω} Ω F (x, u, v)dx ϕ(rn ) ≤ min(p,q) σ n (Kp)1/p +(Kq)1/q ≤ (Kp)1/p + (Kq)1/q min(p,q) (3.3) R Ω sup|s|+|t|<σn F (x, s, t)dx min(p,q) . σn Let γ := lim inf ϕ(r). r→+∞ It follows from (3.2) and (3.3) that γ ≤ lim inf ϕ(rn ) n→+∞ 1/p ≤ (Kp) 1/q + (Kq) min(p,q) R lim Ω n→+∞ sup|s|+|t|<σn F (x, s, t) min(p,q) σn (3.4) min(p,q) = α (Kp)1/p + (Kq)1/q < +∞. From (3.4), it is clear that Λ ⊆]0, γ1 [. For λ ∈ Λ, we claim that the functional Iλ is unbounded from below. Indeed, min(p,q) since λ1 < (Kp)1/p + (Kq)1/q Lβ, we can consider a sequence (τn ) of positive numbers and η > 0 such that τn → +∞ and R min(p,q) F (x, τn , τn )dx 1 B(x0 ,R1 ) < η < L (Kp)1/p + (Kq)1/q , (3.5) q p τn λ + τn p q 6 M. MASSAR, E. M. HSSINI, N. TSOULI EJDE-2012/163 for n large enough. Define a sequence (un ) as follows x ∈ Ω\B(x0 , R2 ) 0, PN 2 i i 2 n [R − (x − x ) ], x ∈ B(x0 , R2 )\B(x0 , R1 ) un (x) = R2τ−R 2 2 0 i=1 2 1 τn , x ∈ B(x0 , R1 ) (3.6) Then (un , un ) ∈ X and 2N τ p π N/2 n (R2N − R1N ), Γ(1 + N/2) R22 − R12 2N τ q π N/2 n kun kqq = (R2N − R1N ). Γ(1 + N/2) R22 − R12 kun kpp = This and (2.1) imply that τp τnq n min(p,q) pL + qL . p q (Kp)1/p + (Kq)1/q 1 Φ(un , un ) = (3.7) By (i1), we have Z Z F (x, un , un )dx ≥ Ψ(un , un ) = F (x, τn , τn )dx. (3.8) B(x0 ,R1 ) Ω Combining (3.5), (3.7) and (3.8), we obtain Iλ (un , un ) = Φ(un , un ) − λΨ(un , un ) Z τp 1 τnq n ≤ + − λ F (x, τn , τn )dx min(p,q) pL qLq p B(x0 ,R1 ) (Kp)1/p + (Kq)1/q Z (3.9) τp τnq 1 n + − λ F (x, τ , τ )dx ≤ n n min(p,q) p q B(x0 ,R1 ) L (Kp)1/p + (Kq)1/q τp τnq 1 − λη n + , < min(p,q) p q L (Kp)1/p + (Kq)1/q for n large enough, so lim Iλ (un , un ) = −∞, n→+∞ and hence the claim follows. The alternative of Theorem 1.2 case (b) assures the existence of unbounded sequence (un ) of critical points of the functional Iλ and the proof of Theorem 2.1 is complete. Proof of Theorem 2.2. First, note that min Φ = Φ(0, 0) = 0. X Let (σn ) be a sequence of positive numbers such that σn → 0+ and R sup|s|+|t|≤σn F (x, s, t)dx = α0 < +∞. lim Ω min(p,q) n→+∞ σn Put min(p,q) σn rn = , δ := lim inf ϕ(r). r→0+ (Kp)1/p + (Kq)1/q (3.10) (3.11) EJDE-2012/163 INFINITELY MANY SOLUTIONS 7 It follows from (3.1) and (3.11) that δ ≤ lim inf ϕ(rn ) n→+∞ 1/p ≤ (Kp) 1/q + (Kq) min(p,q) = α0 (Kp)1/p + (Kq)1/q R lim Ω sup|s|+|t|<σn F (x, s, t) n→+∞ min(p,q) min(p,q) σn (3.12) < +∞. By (3.12), we see that Λ ⊆]0, 1δ [. Now, for λ ∈ Λ, we claim that Iλ has not a local minimum at zero. Indeed, since min(p,q) 0 1 1/p + (Kq)1/q Lβ , we can consider a sequence (τn ) of positive λ < (Kp) numbers and η > 0 such that τn → 0+ and R min(p,q) F (x, τn , τn )dx 1 B(x0 ,R1 ) 1/p 1/q < η < L (Kp) + (Kq) , (3.13) q p τn τ n λ p + q for n large enough. Let (un ) be the sequence defined in (3.6). By combining (3.7), (3.8) and (3.13), and taking into account (i3 ), we obtain Iλ (un , un ) = Φ(un , un ) − λΨ(un , un ) Z τp τnq 1 n + − λ F (x, τn , τn )dx ≤ min(p,q) pL qLq p B(x0 ,R1 ) (Kp)1/p + (Kq)1/q Z τp 1 τnq (3.14) n F (x, τn , τn )dx ≤ min(p,q) p + q − λ B(x0 ,R1 ) L (Kp)1/p + (Kq)1/q τp 1 − λη τnq n < min(p,q) p + q L (Kp)1/p + (Kq)1/q < 0 = Iλ (0, 0) for n large enough. This together with the fact that k(un , un )k → 0 shows that Iλ has not a local minimum at zero, and the claim follows. The alternative of Theorem 1.2 case (c) ensures the existence of sequence (un ) of pairwise distinct critical points (local minima) of Iλ which weakly converges to 0. This completes the proof of Theorem 2.2. Example. It could be possible to consider the same example given in [14] for the p-Laplacian system. Let Ω ⊂ R2 , p = 3, q = 4 and F : R2 → R be a function defined by ( 1 − (an+1 )5 e 1−[(s−an+1 )2 +(t−an+1 )2 ] (s, t) ∈ ∪n≥1 B((an+1 , an+1 ), 1) F (s, t) = 0 otherwise, (3.15) for all x ∈ Ω, where a1 := 2, an+1 := n!(an )5/4 + 2 and B((an+1 , an+1 ), 1) is an open unit ball of center (an+1 , an+1 ). We see that F is non-negative and F ∈ C 1 (R2 ). For every n ∈ N, the restriction of F on B((an+1 , an+1 ), 1) attains its maximum in (an+1 , an+1 ) and F (an+1 , an+1 ) = (an+1 )5 e−1 , 8 M. MASSAR, E. M. HSSINI, N. TSOULI then lim sup F (an+1 , an+1 ) a3n+1 3 n→+∞ + a4n+1 4 EJDE-2012/163 = +∞. So, R F (s, t)dx B(x0 ,R1 ) s3 t4 3 + 4 β : = lim sup s,t→+∞ = |B(x0 , R1 )| lim sup s,t→+∞ F (s, t) 4 = +∞. + t4 s3 3 On the other hand, for every n ∈ N, we have sup |s|+|t|≤an+1 −1 Then lim F (s, t) = a5n e−1 sup|s|+|t|≤an+1 −1 F (s, t) (an+1 − 1)3 n→+∞ and hence for all n ∈ N. = 0, sup|s|+|t|≤σ F (s, t) = 0. σ→+∞ σ3 lim Finally R sup|s|+|t|≤σ F (s, t)dx σ3 sup|s|+|t|≤σ F (s, t) = |Ω| lim inf σ→+∞ σ3 = 0 < Lβ α : = lim inf Ω σ→+∞ So, applying Theorem 2.1, we have that for every λ ∈]0, +∞[ the system ∆(|∆u|∆u) = λFu (u, v) 2 ∆(|∆v| ∆v) = λFv (u, v) u = v = ∆u = ∆v = 0 in Ω, in Ω, (3.16) on ∂Ω, admits an unbounded sequence of weak solutions. References [1] G. A. Afrouzi, S. Heidarkhani; Existence of three solutions for a class of Dirichlet quasilinear elliptic systems involving the (p1 , . . . , pn )-Laplacian, Nonlinear Anal. 70, 135-143 (2009). [2] G. Bonanno, G. Molica Bisci; Infinitely many solutions for a boundary value problem with discontinuous nonlinearities, Bound. Value Probl., 2009 (2009), 1-20. [3] G. Bonanno, G. Molica Bisci; Infinitely many solutions for a Dirichlet problem involving the p-Laplacian, Proc. Royal Soc. Edinburgh, 140A, 737-752 (2010). [4] G. Bonanno, G. Molica Bisci, D. O’Regan; Infinitely many weak solutions for a class of quasilinear elliptic systems, Math. Comput. Model. 52,152-160 (2010). [5] P. Candito, R. Livrea; Infintely many solution for a nonlinear Navier boundary value problem involving p-biharmonic, Studia Univ.”Babeş-Bolyai”, Mathematica, 4 (2010). [6] P. Candito, L. Li, R. Livrea; Infinitely many solutions for a perturbed nonlinear Navier boundary value problem involving the p-biharmonic, Nonlinear Anal., 75 (17) (2012) 63606369. [7] G. D’Agui, A. Sciammetta; Infinitely many solutions to elliptic problems with variable exponent and nonhomogeneous Neumann conditions, Nonlinear Anal., 75, (14) (2012) 5612-5618. [8] H. M. Guo D. Geng; Infinitely many solutions for the Dirichlet problem involving the pbiharmonic-like equation, J. South China Normal Univ. Natur. Sci. Ed., 28 (2009), 18-21. EJDE-2012/163 INFINITELY MANY SOLUTIONS 9 [9] C. Li, C. L. Tang; Three solutions for a Navier boundary value problem involving the pbiharmonic, Nonlinear Anal., 72 (2010), 1339-1347. [10] C. Li, C. L. Tang; Existence of three solutions for (p,q)-biharmonic systems, Nonlinear Anal., 73 (2010), 796-805. [11] B. Ricceri, A general variational principle and some of its applications, J. Comput. Appl.Math., 133 (2000), 401-410. [12] Y. Shen, J. Zhang, Existence of two solutions for a Navier boundary value Problem involving the p-biharmonic, Differential Equations and Applications, 3 (2011), 399-414. [13] Y. Shen, J. Zhang; Multiplicity of positive solutions for a Navier boundary-value problem involving the p-biharmonic with critical exponent, Electronic J. of Differential Equations, 2011, no. 47 (2011) 1-14. [14] D. M. Shoorabi, G. A. Afrouzi; Infinitely many solutions for class of Neumann quasilinear elliptic systems, Bound. Value Probl. 2012, 2012:54 doi:10.1186/1687-2770-2012-54 [15] Y. Wang, Y. Shen; Infinitely many sign-changing solutions for a class of biarmonic equations without symmetry, Nonlinear Anal., 71 (2009), 967-977. [16] W. Wang, P. Zhao; Nonuniformly nonlinear elliptic equations of p-biharmonic type, J. Math. Anal. Appl., 348 (2008) 730-738. [17] Z. Yang, D. Geng, H. Yan; Existence of multiple solutions for a semilinear biharmonic equations with critical exponent, ActaMath. Sci. Ser. A, 27 (1)(2006), 129-142. Mohammed Massar University Mohamed I, Faculty of Sciences, Department of Mathematics, Oujda, Morocco E-mail address: massarmed@hotmail.com El Miloud Hssini University Mohamed I, Faculty of Sciences, Department of Mathematics, Oujda, Morocco E-mail address: hssini1975@yahoo.fr Najib Tsouli University Mohamed I, Faculty of Sciences, Department of Mathematics, Oujda, Morocco E-mail address: tsouli@hotmail.com