Title:Limitations of the Method of Multiple-Time-Scales

Abstract: In the Method of Multiple-Time-Scales (MMTS), the introduction of independent time scales and the elimination of secular terms in the fast time variable, T0 = t, lead to the well-known solvability conditions. Starting from first order, free terms (solutions of the unperturbed equations) emerge in every order in the expansion of the approximate solution. In orders higher than first, the amplitudes of these free terms appear in the solvability conditions. Contrary to the common belief, in the MMTS analysis, these free terms play a role above and beyond the satisfaction of initial conditions: They make feasible mutual consistency among solvability conditions that arise in different orders. In general, this consistency may not be ensured if the free terms are chosen arbitrarily (e.g., set to zero, as is commonly done in many applications). If consistency is not ensured, the analysis may lead to wrong results, or allow only trivial solutions.
The solvability conditions constitute a system of PDE's for the dependence of the amplitudes that appear in the expansion on the slow time scales. However, whenever the free amplitudes must be included to ensure a consistent expansion, these PDE's cannot determine the dependence of the solution on slow time variables beyond the first one, T1=epsilo*t. The dependence on slower time scales, Tn=epsilon^n*t, n > 1, must be imposed either through initial data at, say, T1 = 0, or through requirements on the structure of the approximate solution (based, for example, on physical intuition) that are not related to the validity of the perturbative scheme.
These claims are illustrated through several simple examples, and then discussed in the general case.