The method of Lyapunov functionals for linear difference-differential systems with almost-periodic coefficients (Q5951211)

The stability argument is essentially based on an extension of the Krasovsky method, the basic assumption being that the closeness of initial functions and of the corresponding solutions (trajectories) is `adequately' determined by a norm. It was already due to Lyapunov that the closeness so defined was a very special case. Moreover, the functional derivative, indispensable to analyze the \(V'\), corresponding to the chosen functional \(V\), constitutes an extremely restricted choice, in view of the fact established by Volterra (1913, 1914) that the derivative of a functional contains intrinsically an arbitrary function. A `practically' convenient choice of a norm restricts considerably the scope of the assertions, which are possibly sufficient for most control problems. No plausible evidence is presented that it is indeed so in a wider context. No justification is provided that the singular, critical and exceptional cases of the general Lyapunov theory play no role in other physically significant problems.