Full range of infinite point blow-up exponents for the critical generalized KdV equation
Nailya Manatova
TL;DR
This work proves that the mass-critical gKdV equation admits finite-time blow-up with an infinite-point concentration for every blow-up exponent ν in (1/2,1). By inserting a carefully designed right-hand tail f with decay θ>1 into a refined soliton framework, and by a novel mixed virial-energy functional tailored to the scaling of the problem, the authors construct solutions closely following a modulated Q_b profile. The modulation laws yield λ(s)~s^{−β}, σ(s)~s^{1−β} with β=2(θ−1)/(2θ−1), which in turn produce the blow-up rate ∥∂_xU(t)∥_{L^2}~(T−t)^{−ν} for ν=β/(3β−1)∈(1/2,1). The result extends previous exotic blow-up constructions by covering a full continuum of rates above the critical value and highlights the tail–soliton interaction as the mechanism driving infinite-point blow-up; the method hinges on a refined profile Q_b, a scaling-aware bootstrap, and a monotone Lyapunov functional to control the error terms. Overall, the paper completes the picture of non-generic finite-time blow-up for the L^2-critical gKdV by demonstrating a continuum of infinite-point blow-up exponents and clarifying the role of the tail in shaping the dynamics.
Abstract
For the quintic, mass critical generalized Korteweg-de Vries equation, for any $ν\in (\frac{1}{2}, 1)$, we prove the existence of solutions in the energy space that blow up in finite time $T>0$ with the blow-up rate $\|\partial_x u(t)\|_{L^2} \sim (T-t)^{-ν}$ (infinite point blow-up). These solutions are constructed arbitrarily close to the family of solitons and correspond to the concentration of a soliton traveling at $+\infty$ in space as $t\uparrow T$. This complements the previous results obtained in the work of Martel, Merle, Raphaël in 2015 on infinite point exotic blow-up, which were valid under the technical restriction $ν>\frac {11}{13}$. The value $ν=\frac 12$ corresponds to a critical case to be treated elsewhere. At the technical level, we implement a modification of the virial-energy functional, to allow all $ν> \frac 12$ and simplify the proof of energy estimates.
