What is relative equilibria and flow of PDE?

170 Views Asked by At

Picture below is from Stability of semiclassical bound states of nonlinear Schrödinger equations with potentials. The NLS is nonlinear Schrodinger equation $$ ih\frac{\partial \psi}{\partial t} = -\frac{h^2}{2}\Delta \psi + V\psi-|\psi|^{p-1}\psi $$ What is the flow of a equation ? And what is relative equilibria ?

enter image description here

1

There are 1 best solutions below

0
On BEST ANSWER

The key to understanding a relative equilibrium is the group action. Since you have a group action, you can consider the orbits that its action generates. A relative equilibrium is then a solution to the equations that evolves through such an orbit. Since you didn't include in your image the bound states constructed in the text, I can't specifically point to how this is done in your text. If I had to guess, though, I would say that the bound states change in time precisely through the $S^1$ action, i.e. the time dependence is just phase change: something like $\phi(x,t) = \psi(x) e^{i \omega t}$.

Here "the flow of NLS" is just referring to the time evolution of the solution. The notation / terminology is borrowed from ODE / dynamical systems. In that case for linear problems $\dot{x} = A(t) x$ we have a fundamental matrix (solution operator) $\Phi(t)$, and the solution is given by $x(t) = \Phi(t) x_0$. The "flow" of problem is then the map $t \mapsto x(t) = \Phi(t) x_0$, which tells us how the solutions "flows in time." For linear evolution PDEs we can often do similar things, writing $t \mapsto u(\cdot,t) = S(t) u_0$ for the flow, with $S(t)$ the "solution operator." For example, for the linear heat or Schrodinger equations (without potential), the operator $S(t)$ can be explicitly written out via convolution . This is what the author of the text means here. The question is whether the bound states are stable in the sense that if we start with some data $\phi_0$ near (in some sense) the bound state, does the solution $\phi(\cdot,t)$ stay near (in some sense) the evolution of the bound state for all $t > 0$? This is just the idea of stability from ODE ported to the PDE setting.