Linking methods in critical point theory (Q4223582)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1235895
Language Label Description Also known as
default for all languages
No label defined
    English
    Linking methods in critical point theory
    scientific article; zbMATH DE number 1235895

      Statements

      0 references
      5 January 1999
      0 references
      resonance
      0 references
      linking subsets
      0 references
      existence of critical points
      0 references
      saddle points
      0 references
      Landesman-Lazer resonance
      0 references
      Fučík spectrum
      0 references
      multiplicity of solutions
      0 references
      Linking methods in critical point theory (English)
      0 references
      A vast literature has arisen during the past three decades concerning the existence of critical points \(u\) of multidimensional nonlinear \(C^1\)-functionals \(J: E\to\mathbb{R}\) on a Banach space \(E\), i.e., \(J(u)= c\), \(J'(u)= 0\) for a real constant \(c\) (critical value). The usual approach is to seek sufficient conditions for the existence of a Palais-Smale (PS) sequence \(\{u_n\}\) in \(E\), i.e., \[ J(u_n)\to c,\quad J'(u_n)\to 0\quad\text{as }n\to\infty,\tag{1} \] and then prove that every such sequence has a (strongly) convergent subsequence. The present monograph emphasizes alternatives for (1) of type \[ J(u_n)\to c,\quad J'(u_n)/h(\| u_n\|)\to 0,\tag{2} \] in particular, for suitable monotone functions \(h: \mathbb{R}_+\to \mathbb{R}_+\). One advantage of (2) or variants is that they can be weaker than (1), for example if \(h(t)= (t+ 1)^b\), \(b>0\), while the proof of existence of a convergent subsequence of \(\{u_n\}\) is not more difficult than for (1). Another scenario arises if \(h\) is a suitable decreasing function: A sequence satisfying (2) may have a convergent subsequence while a PS-sequence does not. A principal theme of the book is development of a systematic way to establish existence of critical points, including saddle points, via the concept of ``linking subsets'' of \(E\). Consider the set \(S\) of mappings \(g_t\in C(E\times [0,1], E)\) such that (a) \(g_t\) is a homeomorphism of \(E\) onto \(E\) for each \(t\in [0,1)\); (b) \(g^{-1}_t\) is continuous on \(E\times [0,1)\); (c) \(g_0\) is the identity map; and (d) for each \(g_t\in S\), there exists \(u_0\in E\) with \(g_1u= u_0\) for all \(u\in E\) and \(g_tu\to u_0\) as \(t\to 1-\) uniformly on bounded subsets of \(E\). Definition: A subset \(A\) of \(E\) is said to link a subset \(B\) of \(E\) if \(A\), \(B\) are disjoint and, for every \(g_t\in S\), there exists \(\tau\in(0, 1]\) such that \(g_\tau A\) intersects \(B\). The hypothesis of linking subsets is exploited throughout in proofs that \(J\) has a critical sequence, i.e., a sequence satisfying (1) or (2) or analogue thereof. A typical result of this type is: If \(A\) links \(B\) and \(\sup_A J\leq \inf_B J= b\), then \(J\) has a PS-sequence with \(b\leq c<\infty\). Many results of similar type, for (1) or (2) or analogues, are contained in Chapters 1, 2, 4, and 5. Also, sufficient conditions for linking and properties of linking are included. Chapters 3-12 contain applications to semilinear boundary value problems in a bounded domain \(\Omega\subset \mathbb{R}^n\), of the type \[ Au= p(x,u)\quad\text{weakly in }\Omega,\quad u\in D(A^{1/2}),\tag{3} \] where \(A\) denotes a selfadjoint linear elliptic operator on \(L^2(\Omega)\) and \(C^\infty_0(\Omega)\subset D(A^{1/2})\subset H^{m,2}(\Omega)\) (Sobolev space). The theory of critical sequences is applied to the functional associated with (3): \[ J(u)= \| A^{1/2} u\|^2_{L^2(\Omega)}- 2 \int_\Omega \int^u_0 p(x, s)ds dx, \] and a proof is provided that the critical sequence has a strongly convergent subsequence in \(D(A^{1/2})\). The difficult case of ``resonance'' is studied in Chapters 3, 5, 8, and 12; this is the case for which \(p(x,t)/t\to \lambda_i=\) eigenvalue of \(A\) as \(| t|\to\infty\), or \(p(x, t)/t\to q_{\pm}(x)\) as \(t\to \pm\infty\) with \(\lambda_i\leq q_{\pm}(x)\leq \lambda_{i+1}\) (double resonance). Other types analyzed are called Landesman-Lazer resonance and strong resonance. Additional topics include nonlinear eigenvalues, the Fučík spectrum, jumping nonlinearities, multiplicity of solutions, and many variants of results already mentioned. A lucid preface and the first chapter outline the objectives of the book. A nice feature is a 14 page chapter at the end with comments and historical information. The bibliography contains 322 entries, many of them very recent. The author's original contributions are indicated by 28 of his publications during 1989-1997 or in press.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references