Measures of noncompactness in spaces of regulated functions with application to semilinear measure driven equations (Q2634933)

The authors consider the so-called semilinear measure driven equations, which they write in the form \[ {\mathrm d}x(t)=Ax(t)+f(t,x(t))\,{\mathrm d}g(t),\quad t\in[0,a].\eqno(1) \] The solution takes values in a Banach space \(X\), \(A:X\to X\) is the infinitesimal generator of a \(C_0\)-semigroup \(T(t)\), \(t\geq 0\), \(g:[0,a]\to\mathbb R\) is nondecreasing and left-continuous, and \(f:[0,a]\times X\to X\). Additionally, it is assumed that \(x\) satisfies the nonlocal condition \[ x(0)=p(x),\eqno(2) \] where \(p\) is a mapping from the space of regulated functions \(G([0,a],X)\) to \(X\). A function \(x:[0,a]\to X\) is called a mild solution of the problem (1)--(2) if it satisfies \[ x(t)=T(t)p(x)+\int_0^t T(t-s)f(s,x(s))\,{\mathrm d}g(s),\quad t\in[0,a]. \] The main result of the paper is a theorem guaranteeing the existence of a mild solution to the problem (1)--(2), provided that \(f\), \(T\), \(p\) satisfy some additional technical conditions. One of them involves the measure of noncompactness \(\alpha\), and requires the existence of an integrable function \(L:[0,a]\to\mathbb R\) such that \(\alpha(f(t,B))\leq L(t)\alpha(B)\) for every bounded set \(B\subseteq X\). Reviewer's remarks: 1) The main results of the paper also include two theorems dealing with the measure of noncompactness in the space of regulated functions. However, Theorems 3.1 and 3.2 can already be found in the authors' earlier paper [``On existence of nonlinear measure driven equations involving non-absolutely convergent integrals'', Nonlinear Analysis: Hybrid Systems 20, 72--81 (2016)]. Theorem 3.1 including its proof is a verbatim copy of Lemma 2.8 from the earlier paper. The statement of Theorem 3.2 is identical to Theorem 2.9 from the earlier paper, but the proof is different; it is unclear why the authors decided to reprove it. 2) The authors freely switch between Lebesgue-Stieltjes integrals \(\int_a^b f\,{\mathrm d}\mu\) and Kurzweil-Stieltjes integrals \(\int_a^b f\,{\mathrm d}g\), saying that the latter are more general. However, some care is needed here. First, one has to explain the relation between \(\mu\) and \(g\), and clearly indicate whether the integration domain in \(\int_a^b f\,{\mathrm d}\mu\) is \([a,b)\) or \((a,b]\). Second, although it is true that Lebesgue-Stieltjes integrability implies Kurzweil-Stieltjes integrability, the equality \(\int_a^b f\,{\mathrm d}\mu=\int_a^b f\,{\mathrm d}g\) need not hold in general. Fortunately, it holds if \(g\) is left-continuous (as assumed in the paper) and if \(\int_a^b f\,{\mathrm d}\mu\) is interpreted as \(\int_{[a,b)} f\,{\mathrm d}\mu\).