formulation to a given problem. In particular it is shown how the use of a convolution bilinear
form makes possible a variational formulation for linear initial value problems. A critical
survey of the three main methods that was devised in the past to solve the same problem is
done.