Periodic solutions of Volterra integral equations (Q1111120)

The author considers the existence and uniqueness of periodic solutions of the equation \(x(t)=f(t)+\int^{t}_{-\infty}k(t,s)x(s)ds\) under the assumptions that \(f(t+T)=f(t)\) and \(k(t+T,s+T)=k(t,s).\) The basic idea is to assume that the resolvent kernel r defined to be the solution of the equation \(r(t,s)=-k(t,s)+\int^{t}_{s}k(t,u)r(u,s)du\) satisfies \(\sup_{t\geq 0}\int^{t}_{0}| r(t,s)| ds<\infty\) and use an argument involving limit equations. A result involving contraction mappings that can be applied to the nonlinear equation \(x(t)=f(t)+\int^{t}_{-\infty}k(t,s)g(x(s))ds,\) is given as well. (The claim in Theorem 4(i) is not correct: take \(a(t)=-2e^{-t}\) and then \(b(t)=2e^{-3t}.)\)