1
$\begingroup$

Given some non zero constant vector $\mathbf{W}$, how do I solve:

$$ \dot{\mathbf{X}} = \mathbf{W}\times \mathbf{X} $$

By imagining the movement of $\mathbf{X}$ I think the solution should be a cosine / sinus, but I can't find a good explanation for this, and perhaps the $\mathbf{W}$ will affect the coefficients of the cosines of $\mathbf{X}$.

I tried going through the definition of a cross product via indices and I wrote:

$$ \dot{X}^i = \epsilon_{mki}W^m X^k $$

But I don't know how to solve this either.

$\endgroup$
2
  • $\begingroup$ My first two thoughts: 1. Break $X$ up into $X_w + X_\bot$ with $X_w$ parallel to $W$ and $X_\bot$ perpendicular to $W$. Alternatively, try: 2. First solve it when $W = \langle 1,0,0\rangle$, then use $SO(3)$ invariance to get the solution for arbitrary $W$. $\endgroup$ Commented Apr 8, 2021 at 16:36
  • $\begingroup$ Notice $X$ is conserved, constraining $\vec{X}$ to a sphere of centre $O$ if $X\ne0$. Orthogonality of $\dot{\vec{X}}$ to $\vec{W}$forces $\vec{X}$ to follow a circle if $W\ne0$. $\endgroup$ Commented Apr 8, 2021 at 17:05

2 Answers 2

4
$\begingroup$

Your idea using indices will help us solve this equation. If we write it out in matrix form, we know that $$\dot{\mathbf{X}}=\left(\begin{matrix}0 &-W^3&W^2 \\ W^3&0&-W^1\\ -W^2 & W^1&0\end{matrix}\right)\mathbf{X}\equiv \mathbf{M}\mathbf{X}.$$

This can be solved using regular ODE methods, by finding the eigenvalues $m_i$ and eigenvectors $M_i$ of $\mathbf{M}$. One of the eigenvectors is clearly $M_1=\mathbf{W}$, with eigenvalue $0$. The other two eigenvalues are $\pm i\sqrt{(W^1)^2+(W^2)^2+(W^3)^2}$, and you can calculate the eigenvectors $M_2$ and $M_3$ yourself. So if you break up your initial vector into a linear combination of the eigenvectors, you will indeed see sinusoidal time evolution of the components perpendicular to $\mathbf{W}$: an initial vector $$\mathbf{X}(0)=p_1 M_1+p_2 M_2+ p_3 M_3=p_1 \mathbf{W}+p_2 M_2+ p_3 M_3$$evolves into the vector $$\mathbf{X}(t)=p_1 \mathbf{W}+p_2 M_2 e^{i t \sqrt{(W^1)^2+(W^2)^2+(W^3)^2}}+ p_3 M_3 e^{-i t \sqrt{(W^1)^2+(W^2)^2+(W^3)^2}}.$$

Edit: In summary, the vector $\mathbf{X}$ will precess around the direction given by $\mathbf{W}$ at frequency $|\mathbf{W}|=\sqrt{(W^1)^2+(W^2)^2+(W^3)^2}$.

$\endgroup$
1
$\begingroup$

Write out the system you have hidden behind that tensor notation.

\begin{align*} \dot{x}_1 &= \phantom{w_0 x_1} -w_3 x_2 + w_2 x_3 \\ \dot{x}_2 &= w_3 x_1 \phantom{{}-w_0 x_2} - w_1 x_3 \\ \dot{x}_3 &= -w_2 x_1 + w_1 x_2 \end{align*}

So $$ \dot{X} = \begin{pmatrix} 0 & -1 & 1 \\ 1 & 0 & -1 \\ -1 & 1 & 0 \end{pmatrix} X \text{.} $$

Then the usual ODE method of diagonalization and extraction of eigenvectors will get solutions. (Random example.)

For instance, where $C$ is a (column) vector of three arbitrary constants of integration, $$ x_1(t) = \frac{1}{|W|^2} \left( w_1(W \cdot C) + (w_2^2+w_3^2, -w_1 w_2, -w_1 w_3) \cdot C) \cos(t |W|) + |W|(0, -w_3, w_2)\cdot C \sin(t |W|) \right) $$

The solutions for $x_2$ and $x_3$ are similar.

$\endgroup$

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.