No label defined (Q4270891)

The author studies the problem of constructing an asymptotic expansion for solutions to the singularly perturbed boundary value problem \[ \mu \dot z=F(x,t), \quad \dot y=f(x,t), \quad \Psi(x(0),x(T),\mu)=0, \] with \(x=(z;y)\), \(\epsilon \) is a small parameter, and the discontinuous vector functions \(F\) and \(f\) are defined as follows: \(F=F_j(x,t),\;f=f_j(x,t)\) for \(j=\text{sign} \Phi(x,t)\), where the functions \(F_j,\;f_j,\;j=\pm 1\), are continuous and \(\Phi\) is a scalar function whose zeroes determine the switching surface \(\Gamma\). It is assumed that there exists a solution to degenerate problem \(0=F,\;\dot y=f,\;y|_{t=0}=y_0\), and this solution intersects \(\Gamma\) at a time \(t_0^1\in(0,T)\). In such a situation, the boundary layer phenomena may arise not only near the ends of the interval \((0,T)\) but also near the point \(t_0^1\). The author shows what an effect this circumstance has on the nature of the asymptotic expansion for the solution \(x(t,\mu)\). An estimate for the remainder of such expansion is obtained.