Talk:Multiple-scale analysis

Hamiltonian bound

I made a small edit and someone reverted it, so this is worth a discussion. The Hamiltonian for the Duffing eq. is H = p^2/2 + q^2/2 + (eps/4) q^4. This is clearly not symmetric in q and p, but the stated bounds on p and q are. So this confused me. The bound on p is right: to maximize p set q = 0, then p_max = sqrt(2 H0) with H0 the initial value of H. The same logic applies for q, but we then have q_max^2/2 + (eps/4) q_max^4 = H0. This gives a more complicated, and smaller, q_max than sqrt(2 H0) as written on the page. Yes, it is true that q < sqrt(2 H0), which we can see by dropping the eps term in the q_max eq. while still keeping it in H0. So a lower q_max exists, which reflects the asymmetry between q and p.

I realize the article is about multiple-scale analysis. The main point that q and p are both bounded by the Hamiltonian nature of the system. I got tripped up by H being asymmetric in q and p, but the bounds being symmetric. So it's worth a comment. Though my original one wasn't clear. I'll try again. Dstrozzi (talk) 13:03, 27 March 2025 (UTC)

The title does not cover the fast-slow systems

The topic is successfully explained here (although lacks rigor). However, the topic "fast-slow" systems, known by the same name (Multiple time scales), are not covered here. From fast-slow systems I mean like the "boundary layers in time", "Fenichel's theory" or "canard-cycles" etc.

So I propose to include this topic here as well. However, this may require reorganizing the entry into two main sections. One for the existing (weakly nonlinear problems + secularity conditions etc.) and a fast-slow systems (singularly perturbed problems + geometric singular perturbation theory). Panthades (talk) 12:46, 3 September 2025 (UTC)