Nonlinear Stokes phenomena in first or second order differential equations
Abstract
We study singularity formation in nonlinear differential equations of order m≤slant 2, y(m)=A(x-1,y). We assume A is analytic at (0,0) and ∂y A(0,0)=λ 0 (say, λ=(-1)m). If m=1 we assume A(0,·) is meromorphic and nonlinear. If m=2, we assume A(0,·) is analytic except for isolated singularities, and also that ∫s0∞ |(s)|-1/2d|s|<∞ along some path avoiding the zeros and singularities of , where (s)=∫0s A(0,τ)dτ. Let Hα=\z:|z|>a>0,(z)∈ (-α,α)\. If the Stokes constant S+ associated to + is nonzero, we show that all y such that x +∞y(x)=0 are singular at 2π i-quasiperiodic arrays of points near i+. The array location determines and is determined by S+. Such settings include the Painlev\'e equations PI and PII. If S+=0, then there is exactly one solution y0 without singularities in H2π-ε, and y0 is entire iff y0=A(z,0) 0. The singularities of y(x) mirror the singularities of the Borel transform of its asymptotic expansion, By, a nonlinear analog of Stokes phenomena. If m=1 and A is a nonlinear polynomial with A(z,0) 0 a similar conclusion holds even if A(0,·) is linear. This follows from the property that if f is superexponentially small along + and analytic in Hπ, then f is superexponentially unbounded in Hπ, a consequence of decay estimates of Laplace transforms.
Turn this paper into a full lesson
ArcXiv compiles a staged curriculum from this paper: 8-12 lessons across beginner → advanced, synthesised section guides, visuals, flashcards, a quiz, exercises, and on-demand deep dives per section. Grounded in the abstract, never invented.