Add n- 1 column a(n- 1) of determinant by (-r(n- 1)) to the last column. +r(n- 1)*a(n- 1))

|A|=|(a 1.an)(ai is the I-th vector of A)

By reducing to absurdity, a2, the determinant value is 0,,,,

Multiply the second column a2 of the determinant by (-r2) to the last column, then one of the columns must be represented by the residual linearity r 1*a 1+r2*a2+.

The last column of the determinant becomes 0. Answer (n- 1). ，a2。 .. +r(n- 1)*a(n- 1), assuming that an can be expressed by residual linearity. The column vectors of invertible matrices must be linearly independent.

Let A be an invertible N- catastrophe matrix ...+R (n-1) * A (n-1) |

Add 0 times (-r 1) from the first column of determinant a65438 to the last column. Linear correlation. . ，r 1*a 1+r2*a2+，。 ? ,

. Let a = (a 1, a2...a (n- 1), then |A| is not equal to 0? ,.，a2。 , so

A=(a 1, if a 1