Nonzero solutions of perturbed Hammerstein integral equations with deviated arguments and applications (Q519286): Difference between revisions

In this extremely well written paper the authors consider the existence of solution to the perturbed Hammerstein integral equation \[ u(t)=\gamma(t)\alpha[u]+\int_a^bk(t,s)g(s)f\Big(s,u(s),u\big(\sigma(s)\big)\Big)\;ds. \] Here \(\alpha[u]\) is a functional realized as \[ \alpha[u]:=\int_a^bu(s)\;dA(s), \] where \(A\) is a function of bounded variation but not necessarily monotone. This means that the measure associated to the Stieltjes integral may change sign. Broadly speaking, the interest in solutions to the perturbed Hammerstein equation is that they can be related to various boundary value problems with nonlocal boundary conditions. And this is precisely the aim of the authors in this paper. The assumptions imposed on the problem are lengthy, but all quite reasonable and natural -- their length much more a consequence of numerous functions and functionals to control than any artificiality or contrivance. Moreover, the authors work in the cone \[ \mathcal{K}:=\left\{u\in\mathcal{C}\big([a,b]\big)\;: \;\min_{t\in\left[\widehat{a},\widehat{b}\right]}u(t)\geq c\| u\|\text{, }\alpha[u]\geq0\right\}, \] where \(\left[\widehat{a},\widehat{b}\right]\) is a subset of \([a,b]\) on which the Harnack-like inequality in the cone must hold. (It should be noted that the numbers \(\widehat{a}\) and \(\widehat{b}\) are determined in terms of initial data in the conditions imposed on the problem, but the statement of these is omitted in this review.) It is interesting to note that, as the cone suggests, solutions to the authors' problem are thus allowed to potentially change sign on \([a,b]\) -- i.e., the solutions are not necessarily positive. With this in hand and by means of the classical fixed point index, the authors then deduce a number of existence results. And they apply certain of these results to some specific boundary value problems, such as the reflection problem \[ u'(t)=h\big(t,u(t),u(-t)\big), \] for \(t\in[-T,T]\), subject to \[ u(-T)-u(T)=\alpha[u]. \] The authors also give a very nice analysis of a problem in thermostat modeling. All in all, the paper is excellently written, very mathematically precise, and yet entirely clear and easy-to-read. It is rich in ideas, and specialists in boundary value problems will certainly find something of value here. But, more generally, those researchers with an interest differential equations will find this worthy of their time. This is also highly recommended for those readers interested in an introduction to the methods and techniques of nonlocal boundary value problem analysis.