Chapter 7
By the end of this section, you should be able to understand:
Notation. For the next set of lectures on linear algebra, $\BF$ stands for $\R$ or $\C.$ Thus, if a statement holds for or applies to both number sets, we may simply state it for $\BF.$ Elements of $\BF$ are often called scalars.
Let $V$ be a nonempty set on which are defined operations "$+$" (called addition) and "$\cdot$" (called scalar multiplication).
$V$ is a vector space (over $\BF$) if the following hold for all $\bfu,\bfv,\bfw\in V$ and all $k,\ell\in\BF$:
(V1) | $\bfu+\bfv\in V$ (closure) |
(V2) | $\bfu+\bfv=\bfv+\bfu$ (additive commutativity) |
(V3) | $\bfu+(\bfv+\bfw)=(\bfu+\bfv)+\bfw$ (additive associativity) |
$V$ is a vector space (over $\BF$) if the following hold for all $\bfu,\bfv,\bfw\in V$ and all $k,\ell\in\BF$:
(V1) | $\bfu+\bfv\in V$ (closure) |
(V2) | $\bfu+\bfv=\bfv+\bfu$ (additive commutativity) |
(V3) | $\bfu+(\bfv+\bfw)=(\bfu+\bfv)+\bfw$ (additive associativity) |
(V4) | $\exists\,{\bf 0}\in V$ such that $\bfu+{\bf 0}=\bfu$ (zero vector, or additive identity) |
(V5) | For each $\bfu\in V$, $\exists\,(-\bfu)\in V$ such that $\bfu+(-\bfu)={\bf 0}$ (additive inverse) |
(V6) | $k\cdot\bfu\in V$ (closure) |
(V7) | $k\cdot(\bfu+\bfv)=k\cdot\bfu+k\cdot\bfv$ (multiplicative-additive distributivity) |
(V8) | $(k+\ell)\cdot\bfu=k\cdot\bfu+\ell\cdot\bfu$ (additive-multiplicative distributivity) |
(V9) | $k\cdot(\ell\cdot\bfu)=(k\ell)\cdot\bfu$ (multiplicative-multiplicative distributivity) |
(V10) | $1\cdot\bfu=\bfu$ (multiplicative identity) |
(V1) | $\bfu+\bfv\in V$ (closure) |
(V2) | $\bfu+\bfv=\bfv+\bfu$ (additive commutativity) |
(V3) | $\bfu+(\bfv+\bfw)=(\bfu+\bfv)+\bfw$ (additive associativity) |
(V4) | $\exists\,{\bf 0}\in V$ such that $\bfu+{\bf 0}=\bfu$ (zero vector, or additive identity) |
(V5) | For each $\bfu\in V$, $\exists\,(-\bfu)\in V$ such that $\bfu+(-\bfu)={\bf 0}$ (additive inverse) |
(V6) | $k\cdot\bfu\in V$ (closure) |
(V7) | $k\cdot(\bfu+\bfv)=k\cdot\bfu+k\cdot\bfv$ (multiplicative-additive distributivity) |
(V8) | $(k+\ell)\cdot\bfu=k\cdot\bfu+\ell\cdot\bfu$ (additive-multiplicative distributivity) |
(V9) | $k\cdot(\ell\cdot\bfu)=(k\ell)\cdot\bfu$ (multiplicative-multiplicative distributivity) |
(V10) | $1\cdot\bfu=\bfu$ (multiplicative identity) |
The scalar multiplication symbol is often omitted. Elements of a vector space are usually called vectors.
1) Identify elements of the set: $$\BF^n = \big\{ \u = \left( u_1, u_2, \ldots , u_n\right)~|~ u_1, u_2, \ldots , u_n \in \BF \big\}$$
2) & 3) Check for closure of addition and scalar multiplication: $$\u+ \v = \left( u_1+v_1, u_2+v_2, \ldots , u_n+v_n\right)$$ $$\;\,k \cdot \u = \left( ku_1, ku_2, \ldots , ku_n\right)$$
4) Identify the vector zero: $\mathbf 0 = \left( 0, 0, \ldots , 0\right)$
5) Identify the inverse additive: $ - \mathbf u = \left( -u_1, -u_2, \ldots , -u_n\right)$
1) Identify elements of the set: $$\BF^n = \big\{ \u = \left( u_1, u_2, \ldots , u_n\right)~|~ u_1, u_2, \ldots , u_n \in \BF \big\}$$
2) & 3) Check for closure of addition and scalar multiplication: $$\u+ \v = \left( u_1+v_1, u_2+v_2, \ldots , u_n+v_n\right)$$ $$\;\,k \cdot \u = \left( ku_1, ku_2, \ldots , ku_n\right)$$
4) Identify the vector zero: $\mathbf 0 = \left( 0, 0, \ldots , 0\right)$
5) Identify the inverse additive: $ - \mathbf u = \left( -u_1, -u_2, \ldots , -u_n\right)$
Now you can continue checking that the other properties hold. 📝
Remark: This is just an strategy you can use. But if you prefer, you can verify each property one by one in the given order, that is, from (V1) to (V10).
1) $M_{m,n} = \left\{ \left( \begin{array}{ccc} a_{11} & \cdots & a_{1n} \\ \vdots & \ddots & \vdots \\ a_{m1} & \cdots & a_{mn} \\ \end{array} \right) ~\Bigg|~ a_{ij}\in \BF, 1\leq i\leq m, 1\leq j\leq n \right\}$
2) & 3) Usual addition and scalar multiplication for matrices.
4) $\mathbf 0 = \left( \begin{array}{ccc} 0 & \cdots & 0 \\ \vdots & \ddots & \vdots \\ 0 & \cdots & 0 \\ \end{array} \right)$
5) $- \mathbf u = \left( \begin{array}{ccc} -a_{11} & \cdots & -a_{1n} \\ \vdots & \ddots & \vdots \\ -a_{m1} & \cdots & -a_{mn} \\ \end{array} \right)$
1) $\mathbf f, \mathbf g\in C[a,b]$, often represented as $f(x), g(x).$
2) & 3)
$\left(\,f+g\right)(x) = f(x) + g(x)$
$\qquad \quad \,\, (k \cdot f)(x) = kf(x)$
4) $\mathbf 0 = \,?$ 🤔
5) $ - \mathbf f = -f(x) $
1) $\mathbf p \in P_n(\BF)$, with $\mathbf p = a_0 + a_1x + \cdots + a_n x^n$ and $a_k\in \BF,$ $\forall k.$
2) & 3) Operations similar to Example 7.4.
4) $\mathbf 0 = \,?$ 🤔
5) $ - \mathbf p = -p(x) .$
For example: $y''+p(x) y' + q(x) y = 0 $ ($y=y(x)$).
1) Let $y_1$ and $y_2$ be solutions.
2) Operations similar to Example 7.4.
3) Superposition principle: $y_1+y_2 $ is also a solution.
4) The zero vector is the zero function.
5) Given a solution $y$, $-y$ is also a solution.
Here we lists several concepts with which you should already be familiar.
Here we lists several concepts with which you should already be familiar.
Let $\beta=\{\bfv_1,\ldots,\bfv_n\}$ be a set of vectors in the vector space $V$. $\,\beta$ is a basis for $V$ if
(B1) | $\,\beta$ is linearly independent; |
(B2) | $\,\beta$ spans $V$. |
Note that the notion of a basis is only defined here for finite sets. A nonzero vector space is finite-dimensional if it contains a finite set of vectors that forms a basis. If no such set exists, the vector space is infinite-dimentional.
Let $V$ be a finite-dimensional vector space. The number of vectors in any basis for $V$ is the same, and this number is known as the dimension of $V$.
An ordered basis for a vector space is a basis endowed with a specific order. For some vector spaces, there is a canonical ordered basis, called a standard basis.
For example, for $\R^3$ we have \[ \beta = \left\{ \left( \begin{array}{c} 1 \\ 0\\ 0 \\ \end{array} \right), \left( \begin{array}{c} 0 \\ 1\\ 0 \\ \end{array} \right), \left( \begin{array}{c} 0 \\ 0\\ 1 \\ \end{array} \right) \right\}. \]
For $P_3 \left( \R \right)$ we have \[ \beta = \left\{1, x, x^2, x^3\right\}. \]
Let $\beta=\{\bfv_1,\ldots,\bfv_n\}$ be a set of vectors in the vector space $V$. Then, $\beta$ is a basis for $V$ iff each $\bfw\in V$ can be uniquely expressed as a linear combination of vectors in $\beta$.
Proof: Exercise 📝
Let $\beta=\{\bfv_1,\ldots,\bfv_n\}$ be an ordered basis for the vector space $V$. For $\bfu\in V$, let $a_1,\ldots,a_n$ be (the unique) scalars such that $$ \bfu=\sum_{i=1}^na_i\bfv_i. $$
The coordinate vector of $\bfu$ relative to $\beta$ is given by $$ [\bfu]_\beta = \begin{pmatrix} a_1\\ \vdots\\ a_n \end{pmatrix}. $$ Another common notation for this is $[\bfu]^\beta$.
Let $\beta'$ be another ordered basis for $V$. The coordinate vector of $\bfu$ relative to $\beta'$ is thus denoted by $[\bfu]_{\beta'}$. The transition matrix from $\beta$ to $\beta'$, denoted by $P_{\beta\to\beta'}$, relates the two coordinate vectors of $\bfu$ as $$ [\bfu]_{\beta'}=P_{\beta\to\beta'}[\bfu]_\beta. $$ If $\beta''$ is yet another ordered basis for $V$, then $$ P_{\beta'\to\beta''}P_{\beta\to\beta'}=P_{\beta\to\beta''}\;\, \Ra \;\, P_{\beta'\to\beta}P_{\beta\to\beta'}=P_{\beta\to\beta}=I, $$ where $I$ is the $n\times n$ identity matrix.
To illustrate, let us consider the two ordered bases $\beta=\{1,x\}$ and $\beta'=\{1+x,2x\}$ for $P_1(\BF)$. As the vector (or polynomial) $\bfu=a+bx$ also can be written as $$ \bfu=a(1+x)+\tfrac{1}{2}(b-a)(2x), $$ we have
$$ [\bfu]_\beta=\begin{pmatrix} a\\ b \end{pmatrix} \quad\text{and}\quad [\bfu]_{\beta'}= \begin{pmatrix} a\\ \frac{1}{2}(b-a) \end{pmatrix}. $$
The corresponding transition matrix $P_{\beta\to\beta'}$ is given as follows:
$$ [\bfu]_\beta=\begin{pmatrix} a\\ b \end{pmatrix} \quad\text{and}\quad [\bfu]_{\beta'}= \begin{pmatrix} a\\ \frac{1}{2}(b-a) \end{pmatrix}. $$
The corresponding transition matrix $P_{\beta\to\beta'}$ is given as follows:
\[ \begin{pmatrix} a\\ \frac{1}{2}(b-a) \end{pmatrix} = \begin{pmatrix} 1 & 0\\ -\frac{1}{2} & \frac{1}{2} \end{pmatrix} \begin{pmatrix} a\\ b\end{pmatrix}, \] so \[ P_{\beta\ra \beta'}= \begin{pmatrix} 1 & 0\\ -\frac{1}{2} & \frac{1}{2} \end{pmatrix}. \]
In general we have that \[ P_{\beta\ra \beta'}= \left( \left[ \v_1\right]_{\beta'} ~|~ \left[ \v_2\right]_{\beta'}~|~\cdots ~|~ \left[ \v_n\right]_{\beta'}\right) \] where $\left[ \v_i\right]_{\beta'}$ is the coordinate vector of basis $\v_i$ in $\beta$ relative to the basis $\beta'$.
We can check this for our example:
We have that $\beta = \left\{\v_1, \v_2\right\}= \left\{1, x\right\}$
and
$\qquad \;\;\quad \quad\beta' = \left\{\v_1', \v_2'\right\}= \left\{1+x, 2x\right\}.$
$\beta = \left\{\v_1, \v_2\right\}= \left\{1, x\right\}$ and $\beta' = \left\{\v_1', \v_2'\right\}= \left\{1+x, 2x\right\}.$
$\v_1 = 1$
$= a_1 \v_1' + a_2 \v_2'$
$=a_1\left(1+ x\right)+a_2\left(2x\right)$
$=a_1+ \left(a_1+ 2a_2\right)x.\qquad \qquad \qquad \;$
Comparing coefficients we get $a_1=1$ and $a_1+2a_2 = 0$, that is, $a_2 = -\dfrac{1}{2}.$
Thus $\ds\left[\v_1\right]_{\beta'}=\begin{pmatrix} 1\\ -\frac{1}{2} \end{pmatrix}.$
$\beta = \left\{\v_1, \v_2\right\}= \left\{1, x\right\}$ and $\beta' = \left\{\v_1', \v_2'\right\}= \left\{1+x, 2x\right\}.$
Thus $\ds\left[\v_1\right]_{\beta'}=\begin{pmatrix} 1\\ -\frac{1}{2} \end{pmatrix}.$
Similarly, we have
$\v_2 = x$ $= a_1 \v_1' + a_2 \v_2'$ $=a_1+ \left(a_1+ 2a_2\right)x.$
Comparing coefficients we obtain $a_1=0$ and $a_2 = \dfrac{1}{2},$ $$ \Ra\, \ds\left[\v_2\right]_{\beta'}=\begin{pmatrix} 0\\ \frac{1}{2} \end{pmatrix}. $$
👉 $\;\ds\left[\v_1\right]_{\beta'}=\begin{pmatrix} 1\\ -\dfrac{1}{2} \end{pmatrix}\;\;$ and $\;\;\ds\left[\v_2\right]_{\beta'}=\begin{pmatrix} 0\\ \dfrac{1}{2} \end{pmatrix}$
👉 $\,\ds P_{\beta\ra \beta'} = \left( \left[ \v_1\right]_{\beta'} ~|~ \left[ \v_2\right]_{\beta'} \right) =\begin{pmatrix} 1 & 0 \\ -\dfrac{1}{2} & \dfrac{1}{2}\end{pmatrix}.$