logo

몫공간의 기저와 차원 📂Linear Algebra

몫공간의 기저와 차원

Proposition1

Let VV be a nn-dimensional vector space, and WVW \le V be a k(<n)k (\lt n)-dimensional subspace. Let the basis of WW be {u1,,uk}\left\{ u_{1}, \dots, u_{k} \right\}. Furthermore, let the basis of VV, extended with this basis, be {u1,,uk,uk+1,,un}\left\{ u_{1}, \dots, u_{k}, u_{k+1}, \dots, u_{n} \right\}. Then:

  • {uk+1+W,,un+W}\left\{ u_{k+1} + W, \dots, u_{n} + W \right\} is a basis for the quotient space V/WV/W.

  • dim(V)=dim(V/W)+dim(W)\dim(V) = \dim(V/W) + \dim(W)

Explanation

The result concerning dimension can also be obtained through alternative proofs.

Proof

Let β={uk+1+W,,un+W}\beta = \left\{ u_{k+1} + W, \dots, u_{n} + W \right\}.

  • β\beta is linearly independent.

    Since the zero vector in V/WV/W is WW, we must show that the only solution that satisfies the following equation is ak+1==an=0a_{k+1} = \cdots = a_{n} = 0.

    ak+1(uk+1+W)++an(un+W)=W a_{k+1}(u_{k+1} + W) + \cdots + a_{n}(u_{n} + W) = W

    According to addition defined in V/WV/W,

    ak+1(uk+1+W)++an(un+W)=(ak+1uk+1++anun)+W=W \begin{align*} a_{k+1}(u_{k+1} + W) + \cdots + a_{n}(u_{n} + W) &= (a_{k+1}u_{k+1} + \cdots + a_{n}u_{n}) + W \\ &= W \end{align*}

    Lemma

    (b) For v1,v2Vv_{1}, v_{2} \in V, it is equivalent to v1v2Wv_{1} - v_{2} \in W if v1+W=v2+Wv_{1} + W = v_{2} + W.

    (c) V/WV/W is a vector space, with the zero vector being 0V+W=W0_{V} + W = W. (0V0_{V} is the zero vector of VV.)

    From the lemma above, we obtain:

    (ak+1uk+1++anun)+W=0V+W (a_{k+1}u_{k+1} + \cdots + a_{n}u_{n}) + W = 0_{V} + W     ak+1uk+1++anun0V=ak+1uk+1++anunW(1) \implies a_{k+1}u_{k+1} + \cdots + a_{n}u_{n} - 0_{V} = a_{k+1}u_{k+1} + \cdots + a_{n}u_{n} \in W \tag{1}

    Clearly, uk+1,,unu_{k+1}, \dots, u_{n} is not an element of WW by definition.

    uk+1,,unW(2) u_{k+1}, \dots, u_{n} \notin W \tag{2}

    However, since uk+1,,unu_{k+1}, \dots, u_{n} is linearly independent, (1)(1) and (2)(2) must be simultaneously satisfied, which implies ak+1uk+1++anun=0V=0Wa_{k+1}u_{k+1} + \cdots + a_{n}u_{n} = 0_{V} = 0_{W}. Therefore, we have:

    ak+1==an=0 a_{k+1} = \cdots = a_{n} = 0

  • spanβ=V/W\span \beta = V/W

    For wWw \in W such that w+W=Ww + W = W, for any given vVv \in V:

    v+W=(a1u1++akuk+ak+1uk+1++anun)+W=(ak+1uk+1++anun)+W=(ak+1uk+1+W)++(anun+W) \begin{align*} v + W &= (a_{1}u_{1} + \cdots + a_{k}u_{k} + a_{k+1}u_{k+1} + \cdots + a_{n}u_{n}) + W \\ &= (a_{k+1}u_{k+1} + \cdots + a_{n}u_{n}) + W \\ &= (a_{k+1}u_{k+1} + W) + \cdots + (a_{n}u_{n} + W) \\ \end{align*}

Hence, β\beta is a basis for V/WV/W. Therefore,

dim(V)=dim(V/W)+dim(W)n=(nk)+k \begin{align*} \dim(V) &= \dim(V/W) + \dim(W) \\ n &= (n-k) + k \end{align*}


  1. Stephen H. Friedberg, Linear Algebra (4th Edition, 2002), p58 ↩︎