$\underline{Def}$
A subset $\beta$ of a vector space $V$ is a basis for $V$ if $\beta$ is linear independent & $span(\beta)=V$.
$\underline{Ex.1}$
Since $\phi$ is linear independent & $span(\phi)= \{ 0 \}$, $\phi$ is a basis for $\{ 0 \}$.
$\underline{Ex.2}$
For $i=1, \cdots, n$, let $e_{i} = (0, \cdots, 0, 1, 0, \cdots, 0,) \in \mathbb{F}^{n}$ where 1 is i-th component.
If $a=(a_{1}, \cdots, a_{n}) \in \mathbb{F}^{n}$, then $a=a_{1}e_{1}+ \cdots + a_{n}e_{n}$; thus, $span(\{e_{1}, \cdots, e_{n} \})= \mathbb{F}^{n}$.
Clearly $\{ e_{1}, \cdots, e_{n} \}$ is linear independent.
Thus, $\{ e_{1}, \cdots, e_{n} \}$ is a basis for $\mathbb{F}^{n}$, called the standard basis for $\mathbb{F}^{n}$.
$\underline{Thm}$ (Unique linear combination by a basis)
Let $\beta = \{ u_{1}, \cdots, u_{n} \}$ be a finite subset of a vector space $V$.
Then $\beta$ is a basis for $V$ iff $\forall v \in V,\; \exists (a_{1}, \cdots, a_{n}) \in \mathbb{F}^{n}$ such that $v=a_{1}v_{1}+ \cdots +a_{n}v_{n}$.
$\underline{Proof}$
$(\Rightarrow)$
(Existience)
Since $\beta$ is a basis for $V$, we have $v \in V=span(\beta)$, so that $\exists a_{1}, \cdots, a_{n} \in \mathbb{F}$ such that $v=a_{1}v_{1}+ \cdots +a_{n}v_{n}$.
(Uniqueness) Suppose $(a_{1}^{'}, \cdots, a_{n}^{'}) \in \mathbb{F}^{n}$ and $v=a_{1}^{'}+ \cdots +a_{n}^{'}v_{n}^{'}$.
Then $(a_{1}-a_{1}^{'})v_{1}+ \cdots +(a_{n}-a_{n}^{'})v_{n}=0$.
Since $\beta$ is linear dependent, we get $a_{i}=a_{i}^{'}$ for $i=1, \cdots, n$.
$(\Leftarrow)$
By assuption, $span(\beta)=V$.
Suppose $a_{1}, \cdots, a_{n} \in \mathbb{F}\; \& \; a_{1}v_{1}+ \cdots +a_{n}v_{n}=0$.
Since $0=0v_{1}+ \cdots +0v_{n}$, the uniqueness says $a_{i}=0$ for $i=1, \cdots, n$.
Thus $\beta$ is linear independent.
$\underline{Rmk}$
Let $\beta=\{ u_{1}, \cdots, u_{n} \}$ be a basis for vector space $V$.
$V=span(\beta)$ is linear isomorphic to $\mathbb{F}^{n}$, denoted by $span(\beta) \simeq \mathbb{F}^{n}$.
$\underline{Thm}$
Suppose a vector space $V$ has a finite set $S$ with $span(S)=V$.
Then $\exists \beta \subset S$ such that $\beta$ is a basis for a $V$; hence $V$ has a finite basis $\beta$.
$\underline{Proof}$
If $S=\phi$ or $S=\{ 0\}$, then $V=\{ 0\}$ and $\beta=\phi \subset S$ is a basis for $V$.
So, suppose $S$ has a non-zero vector $u_{1}$. Then $\{ u_{1} \}$ is linear independent.
If $\{ u_{1}, v \}$ is linear dependent, $\forall v \in S \backslash \{ u_{1} \}$, let $\beta=\{ u_{1} \}$.
Otherwise, we choose a $u_{2} \in S \backslash \{u_{1} \}$ so that $\{ u_{1}, u_{2} \}$ is linear independent.
If $\{ u_{1}, u_{2}, v \}$ is linear dependent, $\forall v \in S \backslash \{ u_{1}, u_{2} \}$, let $\beta = \{ u_{1}, u_{2} \}$.
Otherwise, we choose a $u_{3} \in S \backslash \{u_{1}, u_{2} \}$ so that $\{ u_{1}, u_{2}, u_{3} \}$ is linear independent.
Continuing this process, since $S$ is finite, we obtain a linear independent set $\beta = \{ u_{1}, \cdots, u_{k} \} \subset S$ such that $\beta \cup \{ v \}$ is linear dependent $\forall v \in S \backslash \beta$.
We claim that $S \subset span(\beta)$.
Suppose not; i.e. $\exists v \in S \backslash span(\beta)$. Then $\beta \cup \{ v \}$ is linear independent by thm, a contradinction.
So, $S \subset span(\beta)$. Since $span(S)$ is the smallest subspace of $V$ containing $S$, we have $span(S) \subset span(\beta) \subset span(S)$; so $span(\beta)=span(S)=V$.
'수학 > 선형대수학' 카테고리의 다른 글
[선형대수학] 차원 (Dimension) (0) | 2020.08.18 |
---|---|
[선형대수학] 기저 대체 정리 (Replacement Theorem) (1) | 2020.08.18 |
[선형대수학] 선형 종속, 독립의 성질 (Property of Linear Dependence, Independence) (0) | 2020.08.17 |
[선형대수학] 선형 종속 & 독립 (Linear dependence & independence) (0) | 2020.08.17 |
[선형대수학] 스팬 (Span) (0) | 2020.08.17 |