In a study published in Applied Mathematics and Optimization, researchers Yunru Bai, Leszek Gasinski, and Nikolaos S. Papageorgiou proved a multiplicity theorem establishing at least three nontrivial bounded solutions for nonlinear, nonhomogeneous elliptic equations driven by a double phase operator. The investigated boundary value problem incorporates a superlinear reaction alongside Robin boundary conditions on a bounded spatial domain with a twice continuously differentiable boundary. Combining modern variational tools from critical point theory with truncation mechanisms, comparison techniques, and algebraic topological arguments involving critical groups, the authors demonstrate that the differential problem admits two bounded solutions of constant sign—one strictly positive and one strictly negative—together with a third distinct nontrivial solution.
The differential problem is formulated on a bounded domain Omega in N-dimensional Euclidean space and is driven by an operator combining a weighted p-Laplacian with a classical q-Laplacian, with exponents satisfying 1 < q < p < N and p > 2. The weighted p-Laplacian operator is defined as the divergence of a product containing a nonnegative, Lipschitz continuous weight function a(z) and the gradient term to the power p minus two. The conormal derivative associated with the Robin boundary condition weights the normal derivative by the sum of a(z) times the gradient magnitude to the power p minus two and the gradient magnitude to the power q minus two. The overall differential operator is associated with an energy density function vartheta(z, t) defined as a(z)t^p + t^q for all points z in the domain and all nonnegative values of t. The analytical characteristics of the problem depend critically on the behavior of this weight function across the closure of the domain.
When the minimum value of the weight function a across the closure of the domain is strictly positive, the energy density function exhibits balanced growth, remaining trapped between identical powers of t. Under balanced growth, mathematicians can utilize the classical global regularity theory of Lieberman up to the boundary of the domain. This regularity framework provides powerful analytical instruments, including the equivalence of local minimizers in Holder and Sobolev spaces, nonlinear versions of the Hopf maximum principle, and strong comparison principles within standard Lebesgue and Sobolev spaces. However, when the minimum of the weight function reaches zero, the density function exhibits unbalanced growth, meaning that it is trapped between two distinct powers, specifically t^q from below and a constant times (1 + t^p) from above.
Under unbalanced growth, Lieberman global regularity theory up to the boundary is no longer valid, leaving only local regularity properties. Standard Lebesgue and Sobolev spaces do not provide an adequate analytical framework, requiring a transition to generalized Orlicz and Sobolev-Orlicz spaces. Physically, differential operators with unbalanced growth model diffusion-type processes in media where different subdomains exhibit distinct physical characteristics. For instance, they describe composite materials possessing an energy density with q-growth on the set where the weight function vanishes, and an energy density with p-growth where the weight function is strictly positive.
The mathematical study of unbalanced growth problems, also termed double phase problems or problems with (p, q)-growth, originated with foundational works by Marcellini and Zhikov in the calculus of variations and nonlinear elasticity. While recent investigations have established existence and multiplicity of solutions for double phase problems with Dirichlet boundary conditions, studies addressing Robin boundary conditions have lagged behind. Earlier research on unbalanced growth Robin problems examined superlinear reactions only under restrictive conditions, such as Nehari manifold constraints or single existence results.
To establish a rigorous mathematical foundation, the authors formulate the problem within the generalized Lebesgue-Orlicz space L^vartheta, defined as the set of measurable functions whose modular functional integral of vartheta(z, u(z)) over Omega is finite. Equipped with the Luxemburg norm, this space becomes a separable, reflexive, and uniformly convex Banach space, with reflexivity following from the Milman-Pettis theorem. The authors then construct the generalized Sobolev-Orlicz space W^{1, vartheta}, consisting of functions in L^vartheta whose weak gradients also belong to L^vartheta, equipped with the sum of the Luxemburg norms of the function and its gradient. This space embeds continuously and densely into Lebesgue spaces and Sobolev spaces up to the critical Sobolev exponent q* = Nq/(N – q), with compactness holding for all subcritical exponents.
The formulation relies on precise structural hypotheses denoted by H_0 and H_1. Under hypothesis H_0, the weight function is a nontrivial, nonnegative Lipschitz function, the exponents satisfy p > 2 and p/q < 1 + 1/N, and the boundary parameters beta_1, beta_2 and spatial potentials xi_1, xi_2 are essentially bounded, nonnegative, and nontrivial in pairs. The inequality restricting p/q prevents the exponents from being far apart, ensuring that p remains below the critical exponent q* and facilitating local regularity. Under hypothesis H_1, the reaction f is an L^infinity-locally Lipschitz mapping that vanishes at zero and satisfies subcritical growth bounded by a power r strictly between p and q*. Crucially, f is (p – 1)-superlinear at infinity, but it does not need to satisfy the classical Ambrosetti-Rabinowitz condition.
A key technical element in the analysis involves an auxiliary nonlinear eigenvalue problem for the q-Laplacian with Robin boundary conditions. The authors establish that this eigenvalue problem possesses a smallest eigenvalue lambda_1(q) characterized variationally by the infimum of a Rayleigh quotient involving the q-norm of the gradient, domain potential integrals, and boundary integrals over the q-norm of the function. This eigenvalue is strictly positive, isolated, and simple, with corresponding eigenfunctions belonging to C^1 on the domain’s closure and maintaining a fixed sign. This principal eigenvalue provides a quantitative threshold to control the reaction term near the origin, ensuring that the quotient of q times the primitive potential F(z, x) over |x|^q is asymptotically bounded below the principal eigenvalue.
The proof examines the energy functional phi defined on the Sobolev-Orlicz space as the sum of weighted gradient and boundary functionals minus the integral of the reaction potential. The associated nonlinear gradient operator V is proven to be monotone, bounded, continuous, and of type (S)_+, meaning that weakly convergent sequences satisfying an upper limit condition on the dual pairing must converge strongly in norm. The authors introduce positive and negative truncated functionals, phi_+ and phi_-, which replace the reaction with zero for negative or positive arguments. By estimating the modular functional against the Luxemburg norm, the researchers demonstrate that the trivial solution u = 0 is a local minimizer for all three functionals, as the superlinear reaction is dominated near the origin.
Because the standard Ambrosetti-Rabinowitz condition is omitted, establishing compactness requires delicate estimation. The authors demonstrate that the truncated functionals satisfy the Cerami condition, ensuring that sequences with bounded functional values and asymptotically vanishing scaled derivatives admit strongly convergent subsequences. Combining the Cerami condition with the local minimizer property of zero and mountain pass geometry, the authors apply the mountain pass theorem to produce a nontrivial critical point u_0 for phi_+ and a nontrivial critical point v_0 for phi_-. A De Giorgi-Stampacchia truncation argument using level sets where solutions exceed a threshold k proves that any weak solution belongs to L^infinity. The weak maximum principle then confirms that u_0 is strictly positive and v_0 is strictly negative on the domain.
To obtain the third nontrivial solution, the authors apply an algebraic topological approach based on critical groups and Morse theory. They examine the critical groups of the energy functional at isolated critical points, defined through relative singular homology groups with real coefficients. By approximating the reaction with bounded Caratheodory functions and utilizing C^1-continuity properties, the authors prove that the critical groups of the truncated functionals coincide with those of the original functional at u_0 and v_0, showing that their first critical groups are nontrivial and isomorphic to the real line. Evaluating the Morse relation at t = -1 under the assumption that the critical set contains only zero, u_0, and v_0 produces the numerical contradiction 2(-1)^1 + (-1)^0 = 0. This algebraic contradiction demonstrates the existence of a third nontrivial bounded solution, completing the multiplicity theorem.
Subject of Research: Multiplicity of bounded solutions for superlinear double phase Robin problems
Article Title: Superlinear Double Phase Robin Problems
Article References: Bai, Y., Gasiński, L., & Papageorgiou, N. S. (2027). Superlinear Double Phase Robin Problems. Applied Mathematics & Optimization, 95(1), Article 16. https://doi.org/10.1007/s00245-026-10543-8
Image Credits: AI Generated
DOI: 10.1007/s00245-026-10543-8
Keywords: double phase problems, Robin boundary conditions, p-Laplacian, q-Laplacian, unbalanced growth, Orlicz-Sobolev spaces, critical point theory, mountain pass theorem, Cerami condition, critical groups, Morse theory, Stampacchia truncation, nonlinear elliptic equations, variational methods, constant sign solutions, nonhomogeneous operators, Luxemburg norm, generalized Orlicz spaces
Cite Scienmag News
Reid Dalton. (November 22, 2026). Mathematicians Prove Three Bounded Solutions for Superlinear Double Phase Robin Problems. Scienmag. https://scienmag.com/mathematicians-prove-three-bounded-solutions-for-superlinear-double-phase-robin-problems/
Reid Dalton. "Mathematicians Prove Three Bounded Solutions for Superlinear Double Phase Robin Problems." Scienmag, 22 November 2026, https://scienmag.com/mathematicians-prove-three-bounded-solutions-for-superlinear-double-phase-robin-problems/. Accessed 22 November 2026.
Reid Dalton. "Mathematicians Prove Three Bounded Solutions for Superlinear Double Phase Robin Problems." Scienmag. November 22, 2026. https://scienmag.com/mathematicians-prove-three-bounded-solutions-for-superlinear-double-phase-robin-problems/








