Skip to main content
edited body
Source Link
Carlo Beenakker
  • 188.1k
  • 18
  • 448
  • 651

You might want to first carry out a unitary transformation $A(x)\mapsto U(x)A(x)U^\top(x)$, such that all diagonal elements are zero.$^\ast$
Then $A(x)=BC(x)-C(x)B$ with $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$


$^\ast$ This is always possible for a traceless $A$, see for example https://math.stackexchange.com/q/267192/87355
For a $3\times 3$ real matrix the unitary has the two-parameter form $$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix} \begin{pmatrix} 1&0&0\\ 0&\cos\beta(x)&\sin\beta(x)\\ 0&-\sin\beta(x)&\cos\beta(x) \end{pmatrix} .$$ You can solve first for $\alpha(x)$, $$(\partial_x v_x) \cos ^2\alpha+(\partial_yv_y) \sin ^2\alpha+ (\partial_xv_y+\partial_y v_x)\sin\alpha\cos\alpha=0,$$ and then for $\beta(x)$.

$^\ast$ This is always possible for a traceless $A$ (proof).

For a $3\times 3$ matrix the unitary has the two-parameter form $$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix} \begin{pmatrix} 1&0&0\\ 0&\cos\beta(x)&\sin\beta(x)\\ 0&-\sin\beta(x)&\cos\beta(x) \end{pmatrix} .$$ You can solve first for $\alpha(x)$, $$(\partial_1 v_1) \cos ^2\alpha+(\partial_2v_2) \sin ^2\alpha+ (\partial_1v_2+\partial_2 v_1)\sin\alpha\cos\alpha=0,$$ and then for $\beta(x)$.

You might want to first carry out a unitary transformation $A(x)\mapsto U(x)A(x)U^\top(x)$, such that all diagonal elements are zero.$^\ast$
Then $A(x)=BC(x)-C(x)B$ with $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$


$^\ast$ This is always possible for a traceless $A$, see for example https://math.stackexchange.com/q/267192/87355
For a $3\times 3$ real matrix the unitary has the two-parameter form $$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix} \begin{pmatrix} 1&0&0\\ 0&\cos\beta(x)&\sin\beta(x)\\ 0&-\sin\beta(x)&\cos\beta(x) \end{pmatrix} .$$ You can solve first for $\alpha(x)$, $$(\partial_x v_x) \cos ^2\alpha+(\partial_yv_y) \sin ^2\alpha+ (\partial_xv_y+\partial_y v_x)\sin\alpha\cos\alpha=0,$$ and then for $\beta(x)$.

You might want to first carry out a unitary transformation $A(x)\mapsto U(x)A(x)U^\top(x)$, such that all diagonal elements are zero.$^\ast$
Then $A(x)=BC(x)-C(x)B$ with $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$


$^\ast$ This is always possible for a traceless $A$ (proof).

For a $3\times 3$ matrix the unitary has the two-parameter form $$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix} \begin{pmatrix} 1&0&0\\ 0&\cos\beta(x)&\sin\beta(x)\\ 0&-\sin\beta(x)&\cos\beta(x) \end{pmatrix} .$$ You can solve first for $\alpha(x)$, $$(\partial_1 v_1) \cos ^2\alpha+(\partial_2v_2) \sin ^2\alpha+ (\partial_1v_2+\partial_2 v_1)\sin\alpha\cos\alpha=0,$$ and then for $\beta(x)$.

added 284 characters in body
Source Link
Carlo Beenakker
  • 188.1k
  • 18
  • 448
  • 651

You might want to first carry out a ($x$-dependent) unitary transformation of $A(x)$$A(x)\mapsto U(x)A(x)U^\top(x)$, such that all diagonal elements are zero.$^\ast$
Then $A(x)=BC(x)-C(x)B$ with $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$


$^\ast$ This is always possible for a traceless $A$, see for example https://math.stackexchange.com/q/267192/87355
For a $3\times 3$ real matrix the unitary has the singletwo-parameter form $$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix}.$$$$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix} \begin{pmatrix} 1&0&0\\ 0&\cos\beta(x)&\sin\beta(x)\\ 0&-\sin\beta(x)&\cos\beta(x) \end{pmatrix} .$$ You can solve first for $\alpha(x)$, $$(\partial_x v_x) \cos ^2\alpha+(\partial_yv_y) \sin ^2\alpha+ (\partial_xv_y+\partial_y v_x)\sin\alpha\cos\alpha=0,$$ and then for $\beta(x)$.

You might want to first carry out a ($x$-dependent) unitary transformation of $A(x)$, such that all diagonal elements are zero.$^\ast$
Then $A(x)=BC(x)-C(x)B$ with $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$


$^\ast$ This is always possible for a traceless $A$, see for example https://math.stackexchange.com/q/267192/87355
For a $3\times 3$ real matrix the unitary has the single-parameter form $$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix}.$$

You might want to first carry out a unitary transformation $A(x)\mapsto U(x)A(x)U^\top(x)$, such that all diagonal elements are zero.$^\ast$
Then $A(x)=BC(x)-C(x)B$ with $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$


$^\ast$ This is always possible for a traceless $A$, see for example https://math.stackexchange.com/q/267192/87355
For a $3\times 3$ real matrix the unitary has the two-parameter form $$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix} \begin{pmatrix} 1&0&0\\ 0&\cos\beta(x)&\sin\beta(x)\\ 0&-\sin\beta(x)&\cos\beta(x) \end{pmatrix} .$$ You can solve first for $\alpha(x)$, $$(\partial_x v_x) \cos ^2\alpha+(\partial_yv_y) \sin ^2\alpha+ (\partial_xv_y+\partial_y v_x)\sin\alpha\cos\alpha=0,$$ and then for $\beta(x)$.

added 33 characters in body
Source Link
Carlo Beenakker
  • 188.1k
  • 18
  • 448
  • 651

A simple approach isYou might want to first carry out a ($x$-dependent) unitary transformation of $A(x)$, such that all diagonal elements are zero.$^\ast$
Then (This$A(x)=BC(x)-C(x)B$ with $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$


$^\ast$ This is always possible for a traceless $A$, see for example https://math.stackexchange.com/q/267192/87355
For a ) Then, you take$3\times 3$ real matrix the unitary has the single-parameter form $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$$$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix}.$$

A simple approach is to first carry out a ($x$-dependent) unitary transformation of $A(x)$, such that all diagonal elements are zero. (This is always possible, see for example https://math.stackexchange.com/q/267192/87355 ) Then, you take $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$

You might want to first carry out a ($x$-dependent) unitary transformation of $A(x)$, such that all diagonal elements are zero.$^\ast$
Then $A(x)=BC(x)-C(x)B$ with $$B=\begin{pmatrix} 1&0&0\\ 0&2&0\\ 0&0&3 \end{pmatrix},\;\; C_{ij}(x)=\begin{cases} 0&\text{if}\;i=j\\ \frac{A_{ij}(x)}{B_{ii}-B_{jj}}&\text{if}\;\;i\neq j. \end{cases} $$


$^\ast$ This is always possible for a traceless $A$, see for example https://math.stackexchange.com/q/267192/87355
For a $3\times 3$ real matrix the unitary has the single-parameter form $$U(x)=\begin{pmatrix} \cos\alpha(x)&\sin\alpha(x)&0\\ -\sin\alpha(x)&\cos\alpha(x)&0\\ 0&0&1 \end{pmatrix}.$$

Source Link
Carlo Beenakker
  • 188.1k
  • 18
  • 448
  • 651
Loading