Ajzerman's problem (Q1359897): Difference between revisions

This paper deals with Ajzerman's problem [\textit{M. A. Ajzerman}, Usp. Mat. Nauk, n. Ser. 4, No. 4(32), 187-188 (1949; Zbl 0040.19601)], i.e.: If the trivial solution of the system \[ \dot x_1=\sum^n_{k=1} a_{1k}x_k+f(x_1) \dot x_i=\sum^n_{k=1}a_{ik}x_k, \qquad i=2,\dots, n \tag{*} \] is asymptotically stable for a linear function \(f(x_1)=bx_1\) with \(b\) satisfying the condition \(\alpha <b <\beta\), will the trivial solution of the system \((*)\) be stable in the large for a nonlinear function \(f(x_1)\) satisfying the condition \[ \alpha < f(x_1)/x_1 < \beta ? \] It is proved that if \(A=(a_{ik})\) is selfadjoint, \(\alpha > -\infty\) and \(\beta < \infty\), \(b\in B=[\alpha,\beta ]\cup \{a\}\), where \(a=\lim f(x)/x\) as \(x\to 0\) if the limit exists, or \(a=0\) otherwise, Ajzerman's problem is affirmative.