Matrix Theory study notes[6]

transition matrix

1. a basis of linear space $V^k$,which is $x_1,x_2,...x_k$,can be called as a coordinate system.let vector $v\in V^k$ and it can be linear expressed on this basis as $v=a_1{x}_1+a_2{x}_2+...+a_k{x}_k$,the $a_1,a_2,....,a_k$ is coordinate in this coordinate system denoted by $(a_1,a_2,...,a_k{)}^T$.
2. the various coordinate systems for the same vector are different usually because of non-uniqueness of basis of a linear space. for the first basis which is $x_1,x_2,...x_k$ ,the coordinate is $(a_1,a_2,...,a_k{)}^T$ and there are the second basis $x_1^{\prime },x_2^{\prime },...x_k^{\prime }$ to coorespond another coordinate $(a_1^{\prime },a_2^{\prime },...,a_k^{\prime }{)}^T$,also can be explain that $v=a_1{x}_1+a_2{x}_2+...+a_k{x}_k=a_1^{\prime }{x}_1^{\prime }+a_2^{\prime }{x}_2^{\prime }+...+a_k^{\prime }{x}_k^{\prime }$.
3. let $v\in V^k$ and $x_1,x_2,...x_k$ is a basis of linear space,then $v$ can uniquely be separated into the linear combination that $v=a_1{x}_1+a_2{x}_2+...+a_k{x}_k$.
4. transition matrix is useful for the coversion from a basis of linear space to another basis in the same linear space.let $X=x_1,x_2,...,x_k$ is the old basis of $V^k$ and $X^{\prime }=x_1^{\prime },x_2^{\prime },...,x_k^{\prime }$ is the new basis of that,the transtion matrix $P$ is $n\times n$ matrix which column vectors are coordinate representation of vectors in the new basis under old basis $X$.
   $$x_j^{\prime }=\sum _{i=1}^n{p}_{ij}{x}_i(j=1,2,\dots ,n)$$
   that matrix form is as follows.
   $$\begin{gathered} P=\left[\begin{array}{cccc}[x_1^{\prime }{]}_X& [x_2^{\prime }{]}_X& \cdots & [x_n^{\prime }{]}_X\end{array}\right] \\ {x}_1^{\prime }=p_{11}{x}_1+p_{21}{x}_2+...+p_{n1}{x}_n \\ {x}_2^{\prime }=p_{12}{x}_1+p_{22}{x}_2+...+p_{n2}{x}_n \\ .... \\ {x}_n^{\prime }=p_{1n}{x}_1+p_{2n}{x}_2+...+p_{nn}{x}_n \end{gathered}$$
   $$X^{\prime }=XP$$
   $$P=\left[\begin{array}{cccc}{p}_{11}& p_{12}& ...& p_{1n}\\ ...& ...& ...& ...\\ p_{n1}& p_{n2}& ...& p_{nn}\end{array}\right]$$

• if the vector $y$ 's coordinate under the basis $X^{\prime }$ is $[y{]}_{X^{\prime }}$,then its coordinate under the basis $X$ is that $[y{]}_X=P[y{]}_{X^{\prime }}$

• inverse transformation：the transition matrix from $X$ to $X^{\prime }$ is $[y{]}_{X^{\prime }}=P^{-1}[y{]}_X$

• the transition matrix certainly is an inverse matrx.

• if the transition maxtrix P can help to fulfill the conversion from basis $X$ to basis $X^{\prime }$ and the $P^{\prime }$ which is a transition maxtrix can transfer from the basis $X^{\prime }$ to basis $X^{\prime \prime }$,then $P{P}^{\prime }$ helps the basis $X$ convert to the basis $X^{\prime \prime }$

5. the matrix express of linear transformation,which is transformation from the abstract linear mapping to concrete number matrix, can define completely all linear transformation through basical vector’s transformation.

• let $T:V\to W$ is the linear transformation as follows.
  • $V$ is $n$ dimenstion vector space,its basis is B V = { v 1 , … , v n } B_V = \{\mathbf{v}_1, \dots, \mathbf{v}_n\} BV​={v1​,…,vn​}
  • $W$ is $m$ dimenstion vector space，its basis is B W = { w 1 , … , w m } B_W = \{\mathbf{w}_1, \dots, \mathbf{w}_m\} BW​={w1​,…,wm​}
• in the first place,$T$ has an effect on each basis vector ${\mathbf{v}}_j$ to generate the $T({\mathbf{v}}_j)$.in the second place,the $T({\mathbf{v}}_j)$ is expressed as the linear combination of basis $B_W$ of $W$.
  $$T({\mathbf{v}}_j)=a_{1j}{\mathbf{w}}_1+a_{2j}{\mathbf{w}}_2+\cdots +a_{mj}{\mathbf{w}}_m$$
  all coefficient $a_{ij}$ arrange by columns to form matrix $A$.
  $$A=\left[\begin{array}{cccc}{a}_{11}& a_{12}& \cdots & a_{1n}\\ a_{21}& a_{22}& \cdots & a_{2n}\\ ⋮& ⋮& \ddots & ⋮\\ a_{m1}& a_{m2}& \cdots & a_{mn}\end{array}\right]$$

  • the j-th column：the coordinate of $T({\mathbf{v}}_j)$ under the basis $B_W$ .

  • a vector $\mathbf{v}=c_1{\mathbf{v}}_1+\cdots +c_n{\mathbf{v}}_n$ will transform to the following form because of the linearity of $T$.
    $$T(\mathbf{v})=c_{1}T({\mathbf{v}}_1)+\cdots +c_{n}T({\mathbf{v}}_n)$$

    • input coordinate: $[\mathbf{v}{]}_{B_V}=\left[\begin{array}{c}{c}_1\\ ⋮\\ c_n\end{array}\right]$
    • output coordinate: $[T(\mathbf{v}){]}_{B_W}=A[\mathbf{v}{]}_{B_V}$

6. the matrix express of linear transformation bases on the basis,the matrix express can vary with the basis,although those various matrix express illustrate the same linear transformatin,that is to say,those matrix express certainly have Matrix similarity.

• $T:V\to V$ is a linear transformation.
  • B 1 = { v 1 , … , v n } B_1 = \{ \mathbf{v}_1, \dots, \mathbf{v}_n \} B1​={v1​,…,vn​} is a set of bases,cooresponding to the matrix express $A$.
  • B 2 = { w 1 , … , w n } B_2 = \{ \mathbf{w}_1, \dots, \mathbf{w}_n \} B2​={w1​,…,wn​}is another group of bases,cooresponding to the matrix express $B$.
• the transformation matrix from $B_1$ to $B_2$ satisfies the following requirement.
  $${\mathbf{w}}_j=\sum _{i=1}^n{p}_{ij}{\mathbf{v}}_i$$
  that is as follows.
  $$[{\mathbf{w}}_j{]}_{B_1}=P\cdot [{\mathbf{w}}_j{]}_{B_2}$$
  under the basis $B_1$,the matrix express of $T$ is $A$
  $$[T(\mathbf{v}){]}_{B_1}=A[\mathbf{v}{]}_{B_1}$$
  under the basis $B_2$,the matrix express of $T$ is $B$
  $$[T(\mathbf{v}){]}_{B_2}=B[\mathbf{v}{]}_{B_2}$$
• to prove the similar relation.

$$[T(\mathbf{v}){]}_{B_1}=P[T(\mathbf{v}){]}_{B_2}=PB[\mathbf{v}{]}_{B_2}$$
$[\mathbf{v}{]}_{B_1}=P[\mathbf{v}{]}_{B_2}$,so $[T(\mathbf{v}){]}_{B_1}=A[\mathbf{v}{]}_{B_1}=AP[\mathbf{v}{]}_{B_2}$
$$PB=AP\Rightarrow B=P^{-1}AP$$
we continue to exlain that with an example.

let $T:{\mathbb{R}}^2\to {\mathbb{R}}^2$,$T(x,y)=(2x+y,x+2y)$
the matrix under the standard basis B 1 = { e 1 , e 2 } B_1 = \{ \mathbf{e}_1, \mathbf{e}_2 \} B1​={e1​,e2​}

$$T({\mathbf{e}}_1)=(2,1),T({\mathbf{e}}_2)=(1,2)$$
so
$$A=\left[\begin{array}{cc}2& 1\\ 1& 2\end{array}\right]$$

the matrix under the new basis $B_2=\{(1,1),(1,-1)\}$

• transformation matrix $P$：
  $$\begin{gathered} (1,1)=1\cdot {\mathbf{e}}_1+1\cdot {\mathbf{e}}_2 \\ (1,-1)=1\cdot {\mathbf{e}}_1-1\cdot {\mathbf{e}}_2 \\ \Rightarrow P=\left[\begin{array}{cc}1& 1\\ 1& -1\end{array}\right] \end{gathered}$$
• $B=P^{-1}AP$：
  $$P^{-1}=\frac{1}{2}\left[\begin{array}{cc}1& 1\\ 1& -1\end{array}\right]$$
  $$B=\frac{1}{2}\left[\begin{array}{cc}1& 1\\ 1& -1\end{array}\right]\left[\begin{array}{cc}2& 1\\ 1& 2\end{array}\right]\left[\begin{array}{cc}1& 1\\ 1& -1\end{array}\right]=\left[\begin{array}{cc}3& 0\\ 0& 1\end{array}\right]$$

7. to define a linear trasformation that $T:{\mathbb{R}}^2\to {\mathbb{R}}^2$,the standard basis is ${\mathbf{e}}_1=\left(\begin{array}{c}1\\ 0\end{array}\right),{\mathbf{e}}_2=\left(\begin{array}{c}0\\ 1\end{array}\right)$.

let $T(x,y)=(2x+y,x-3y)$：
$$[T]=\left(\begin{array}{cc}2& 1\\ 1& -3\end{array}\right)$$

let $\mathbf{v}=\left(\begin{array}{c}x\\ y\end{array}\right)=x{\mathbf{e}}_1+y{\mathbf{e}}_2$，so
$$T(\mathbf{v})=xT({\mathbf{e}}_1)+yT({\mathbf{e}}_2)=x\left(\begin{array}{c}2\\ 1\end{array}\right)+y\left(\begin{array}{c}1\\ -3\end{array}\right)=\left(\begin{array}{c}2x+y\\ x-3y\end{array}\right)$$
that result is also computed with the following matrix multiplicatoin.
$$[T]\mathbf{v}=\left(\begin{array}{cc}2& 1\\ 1& -3\end{array}\right)\left(\begin{array}{c}x\\ y\end{array}\right)=\left(\begin{array}{c}2x+y\\ x-3y\end{array}\right)$$

references

1. deepseek
2. 《矩阵论》