Multiple positive solutions for a class of integral inclusions (Q2479365): Difference between revisions

The author deals with integral inclusions of the following form \[ x(t) \in f\bigl(t,x(t)\bigr)\int_{0}^{1}k(s)U\bigl(t,s,x(s)\bigr)ds \quad \text{for }\, t \in [0,1],\tag{*} \] where \(f\) and \(k\) are some continuous functions and \(U\) is an \(L^{1}\)-Carathéodory multivalued map. Under suitable assumptions the existence of at least one (Theorem 1) or at least two (Theorem 2) positive solutions to (*) is established. The proofs are based on an expansion and compression fixed point theorem due to \textit{R. P. Agarwal} and \textit{D. O'Regan} [J. Differ. Equations 160, No.~2, 389--403 (2000; Zbl 1008.47055)] for upper semicontinuous \(k\)-set contractive (with \(0\leq k <1\)) multivalued maps defined on a subset of a cone \(P\) in a Banach space, with values in the set of all convex, compact and nonempty subsets of \(P\). Unfortunately, the paper is not easy to read. Some definitions are obscure; e.g. the definition of a \(k\)-set contraction on page 21. If the space \(E\) in question is simply \(\mathbb{R}^{n}\) (see page 20, below the formula (*), then this definition does not make sense. There are also some mistakes. For example, on page 25 there is written: ``For any bounded \(D \subset P\) and any given \(\varepsilon >0\), there exist finitely many balls, say, \(B_{\varepsilon}(x_{i})=\{x: \left\| x-x_{i}\right\| \leq \varepsilon\}\) for \(i=1,2,\ldots,m\), such that \(D \subset \bigcup_{i=1}^{m}B_{\varepsilon}(x_{i})\).'' Here \(\beta\) stands for the Hausdorff measure of noncompactness and \(P\) is a cone in infinite dimensional Banach space, so in general the above statement is not true. And finally, there are some computational mistakes; e.g. in the Example 1 on page 28 there is written \(\eta=\frac{5}{8}\). Unfortunately, in such a case this example fails to illustrate Theorem 1, in which there is assumed that \(0<\eta<\frac{1}{2}\). A similar comment concerns Example 2 from this paper.