Cosine Families of Unbounded Operators (Q3317427)

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 3847838
Language Label Description Also known as
default for all languages
No label defined
    English
    Cosine Families of Unbounded Operators
    scientific article; zbMATH DE number 3847838

      Statements

      Cosine Families of Unbounded Operators (English)
      0 references
      0 references
      1985
      0 references
      unique strong solution
      0 references
      Dunford functional calculus
      0 references
      The abstract initial value problem \(u''(t)+Au(t)=f(t)\), \(t\in {\mathbb{R}}\), \(u(0)=u_ 0\), \(u'(0)=u_ 1\) is considered in an arbitrary Banach space E. Existence of a unique strong solution is proved provided the time independent closed linear operator A satisfies: (1) there exists \(\omega>0\) such that \(\sigma\) (-A) is contained in the unbounded domain left to the parabola \(P_{\omega}=\{\lambda \in {\mathbb{C}}: \lambda =\omega^ 2e^{i\psi}/\cos^ 2(\psi /2)\), \(0\leq | \psi |<\pi \}\), and (2) the resolvent \((\lambda +A)^{-1}\) decays to order \(O(| \lambda |^{-\epsilon}| Im \lambda |^{\epsilon -1})\) at infinity for some \(\epsilon\in [0,1]\). In this case the commonly known framework of strongly continuous cosine families does not work since the requirement of well-posedness is violated. Using the Dunford functional calculus it is shown instead that there exists a cosine family of unbounded linear operators C(t), \(t\in {\mathbb{R}}\) in E which is represented by \[ C(t)v=v-\frac{A}{2\pi i}\int_{P_{\omega}}\cos h(t\sqrt{\lambda})(\lambda +A)^{-1}v\frac{d\lambda}{\lambda}. \] The associated sine family \(S(t)v=\int^{t}_{0}C(s)vds\) is utilized to prove the solution representation \(u(t)=C(t)u_ 0+S(t)u_ 1+\int^{t}_{0}S(t-S)f(s)ds\) for prescribed \(u_ 0\in E_ 2\), \(u_ 1\in E_ 1\) and for appropriate f, where \(E_ j=\{v\in E:C(\cdot)v\in C^ j({\mathbb{R}};E)\}\), \(j=1,2\) and \(D(A^ 2)\subset E_ 2\subset E_ 1\subset D(A)\). If merely \(u_ 0\in D(A)\) and \(u_ 1\in D(A)\), then the above u(t) is to be interpreted as a generalized solution to the abstract initial value problem. Examples of operators A satisfying (1) and (2) are provided. The author extends herewith results obtained in Proc. R. Soc. Edinb., Sect. A 87, 305-318 (1981; Zbl 0453.47023).
      0 references

      Identifiers