由 yijun 在 05 Mar 2007, 18:47
[b]线性空间[/b]
一个线性空间,就是由约定的两个集合、四种运算构筑出来的。
两个集合:
[t]\{X, +\}[/t]为一个Abelian群;
[t]\{A, +,*\}[/t]为一个域;
四种运算:
除了上述群和域中的三种运算之外,
[t]\times: A\times X\rightarrow X[/t] 是这两个集合之间的运算。
( [t]\times: Z\times X \rightarrow X[/t] 这个运算约定已经暗示了线性空间的结构。)
使用这四种运算: +, +, *, [t]\times[/t] ,如果 [t]\{x,y\}\subset X ,\{\alpha,\beta\}\subset A[/t] ,那么我们可以做如下三种基本的运算组合:
[t](\alpha+\beta)\times x[/t]
[t](\alpha*\beta)\times x[/t]
[t]$ \alpha\times(x+y) $[/t]
然后我们进一步约定:
[t]$ (\alpha+\beta)\times x=(\alpha\times x)+(\beta\times x) $[/t]
[t]$ (\alpha*\beta)\times x=\alpha*(\beta\times x) $[/t]
[t]$ \alpha\times(x+y)=(\alpha\times x)+(\beta\times y) $[/t]
这就完备地约定了一个线性空间[t]X[/t]。
然后我们就需要知道,这样一个集合-线性空间,可以具有什么样的构造和属性。
[b]维度和子空间[/b]
for 0 of group [t]$\{X, +\}$, what does $x+y+,...,+z=0$[/t] means?
[t]$\forall x\in X, 0*x=0, 0\in A, 0\in X $[/t];
(x+0*x=(1+0)*x=x)
[t]$\forall\alpha\in A, \alpha*0=0, 0\in X $[/t];
if [t]$\alpha\ne0,x\ne0$,then $\alpha*x\ne0$[/t];
but, [t]$\alpha*x+\beta*y=0$[/t] is possible! when [t]$\exists\alpha,\beta, x, y\ne0$[/t].
then, what does that means?
means that:
[t]$\forall\gamma\in A,\{z|z=\gamma*x\}$[/t] is a linear space; any such [t]$y\in\{z\}$[/t]; x and y is symmetry in the lemma;
[t]$\{Z\}$[/t] is a subgroup of [t]$X$[/t].
if [t]$\exists$[/t] y is not [t]$\in\{z\}$[/t], then
[t]$\forall\alpha,\beta\ne0,\alpha*x+\beta*y\ne0$[/t], at this situation, [t]$\alpha*x+\beta*y+\gamma*z=0$[/t] is possible!
what does this means?
means that:
[t]$\forall\alpha,\beta\in A,\{z|z=\alpha*x+\beta*y\}$[/t] is a linear space; y and z is symmetry in this lemma; and [t]$\{z\}$[/t] is a subgroup of [t]$X$[/t].
let's go on!
if [t]$\{z\}=X$[/t], and z can be expressed as
[t]$z=\alpha*x+...+\beta*y$[/t]
then we can say [t]$\{x,...,y\}$[/t] is bases of [t]$X$[/t]. the number of the elements of [t]$\{x,...,y\}$[/t] is the dimension number of [t]$X$[/t].
because of the form [t]$x+y+,...,+z=0$[/t] can be used to generate all the space [t]$X$[/t], we name such [t]$\{x,y,...z\}$[/t] is linear dependent. and if [t]$z=x+...+y$[/t], we name[t]$x+...+y$[/t] is a linear
combination of z.
theorem
any vector z of [t]$X$[/t] can be expressed as its bases's unique linear combination.
[b]coordinate transform[/b]
if [t]$X$[/t] is a n-dimension linear space, then any non-linear dependent n elements of[t]$X$[/t] can be used as its base.
[b]isomorphism[/b]
How to retain the structure of a linear space?