On the Use of Nonsinusoidal Approximating Functions for Nonlinear Oscillation Problems

When treating nonlinear oscillation problems by the Ritz-Galerkin method the approximating function is conventionally chosen as a sinusoid or a polynomial of sinusoids with their amplitudes as the open parameters. In this paper the advantages are pointed out that may be gained from choosing a nonsinusoidal approximating function that contains additional parameters related to the shape of the function.