Sylvester不等式
设A、B分别是 s × n 、 n × m s\times n、n\times m s×n、n×m,则 r a n k ( A B ) ≥ r a n k ( A ) + r a n k ( B ) − n rank(AB)\ge rank(A)+rank(B)-n rank(AB)≥rank(A)+rank(B)−n
证明
只需证 n + r a n k ( A B ) ≥ r a n k ( A ) + r a n k ( B ) n+rank(AB)\ge rank(A)+rank(B) n+rank(AB)≥rank(A)+rank(B)
n + r a n k ( A B ) = r a n k ( I n 0 0 A B ) n+rank(AB)=rank \begin{pmatrix}I_n & 0\\0 &AB \end{pmatrix} n+rank(AB)=rank(In00AB)作分块矩阵的初等行变换
( I n 0 0 A B ) ⟶ ( I n 0 A A B ) ⟶ ( I n − B A 0 ) ⟶ ( I n B A 0 ) ⟶ ( B I n A 0 ) \begin{pmatrix} I_n&0\\0&AB\end{pmatrix}\longrightarrow \begin{pmatrix}I_n&0\\A&AB\end{pmatrix} \longrightarrow \begin{pmatrix}I_n&-B\\A&0\end{pmatrix}\\ \longrightarrow \begin{pmatrix}I_n&B\\A&0\end{pmatrix} \longrightarrow \begin{pmatrix}B&I_n\\A&0\end{pmatrix} (In00AB)⟶(InA0AB)⟶(InA−B0)⟶(InAB0)⟶(BAIn0)
根据分块矩阵的初等行变换不改变矩阵的秩有：
r a n k ( I n 0 0 A B ) = ( B I n 0 A ) ≥ r a n k ( B ) + r a n k ( A ) rank \begin{pmatrix}I_n&0\\0&AB\end{pmatrix} =\begin{pmatrix}B&I_n\\0&A\end{pmatrix} \ge rank(B)+rank(A) rank(In00AB)=(B0InA)≥rank(B)+rank(A)
因此r a n k ( A B ) ≥ r a n k ( A ) + r a n k ( B ) − n rank(AB)\ge rank(A)+rank(B)-n rank(AB)≥rank(A)+rank(B)−n
8429
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
