On the nonexistence of pure multi-solitons for the quartic gKdV equation

We consider the quartic (nonintegrable) (gKdV) equation. Let u(t) be an outgoing 2-soliton of the equation, i.e. a solution behaving exactly as the sum of two solitons (of speeds c1 and c2) for large positive time. In arXiv:0910.3204, for nearly equal solitons, the solution u(t) is computed up to some order of epsilon=1-c2/c1, everywhere in time and space. In particular, it is deduced that u(t) is not a multi-soliton for large negative time, proving the nonexistence of pure multi-soliton in this context. In the present paper, we prove the same result for an explicit range of speeds: 3/4 c1< c2< c1, by a different approach, which does not longer require a precise description of the solution. In fact, the nonexistence result holds for outgoing N-solitons, for any N>1, under an explicit assumption on the speeds, which is a natural generalization of the condition for N=2.

No reviews yet.

No comments yet.