Affine-periodic solutions for generalized ODEs and other equations (Q2688180): Difference between revisions

Given a real number \(T>0\) and an invertible matrix \(Q\in\mathbb R^{n\times n}\), a solution \(x:\mathbb R\to\mathbb R^n\) of a classical differential equation \(x'(t)=f(t,x(t))\) is called \((Q,T)\)-affine-periodic if it satisfies \(x(t+T)=Qx(t)\) for all \(t\in\mathbb R\). Various authors have studied the existence of \((Q,T)\)-affine-periodic solutions for equations whose right-hand side satisfies \(f(t+T,x)=Qf(t,Q^{-1}x)\). The authors of the present paper try to establish similar results for generalized ordinary differential equations of the form \(\frac{dx}{d\tau}=DF(x,t)\), which are understood in Kurzweil's sense. They obtain some results guaranteeing the existence of \((Q,T)\)-affine-periodic solutions, provided that the right-hand side satisfies \(F(t+T,x)=QF(t,Q^{-1}x)\), and certain other technical assumptions hold. Unfortunately, the last condition seems too strong, and the theory developed in the paper does not cover even the existing theory for ODEs. It is known that a classical ODE corresponds to a generalized ODE with right-hand side \(F(t,x)=\int_0^t f(s,x)\,ds\). The choice \(Q=I\) corresponds to \(T\)-periodic solutions, and the condition for classical ODEs is \(f(t+T,x)=f(t,x)\), but this does not imply \(F(t+T,x)=F(t,x)\) unless \(\int_0^T f(s,x)\,ds=0\); this somewhat artificial condition appears in the last three sections dealing with measure differential equations, impulsive equations, and dynamic equations on time scales.