Prove that det(bb(AB)) = det(bb(A)) det(bb(B)) det(AB)=det(A)det(B)?

2 Answers
Sep 14, 2017

See explanation...

Explanation:

The nn dimensional measure polytope is the nn dimensional analogue of the cube. So the unit measure polytope is a line segment, square, cube, hypercube, etc with sides of length 11.

Suppose MM is any n xx nn×n matrix.

Multiplication by MM is a linear operator from nn dimensional space to itself.

The nn dimensional measure (length, area, volume, hypervolume, etc.) of the image of the unit measure polytope (unit line segment, unit square, unit cube, unit hypercube, etc.) under this linear transform is the determinant of MM.

For example, in 22 dimensional real space, a real 2 xx 22×2 matrix MM with non-zero determinant dd maps the unit square to a parallelogram of area dd.

Since matrix multiplication is linear, applying MM to any nn dimensional object of measure xx will result in an object of measure x det(M)xdet(M)

So if we have two matrices AA, BB and apply one then the other to the unit measure polytope, the measure of the resulting polytope will be det(AB) = det(A)*det(B)det(AB)=det(A)det(B)

Sep 14, 2017

We seek to prove that:

det(bb(AB)) = det(bb(A)) det(bb(B)) det(AB)=det(A)det(B)

Consider, as a specific case, a general 2xx22×2 system

bb(A) = ( (a_11, a_12), (a_21, a_22) ) and bb(B) = ( (b_11, b_12), (b_21, b_22) )

Evaluation of the LHS:

det( bb(A) ) = a_11 a_22 - a_21 a_12
det( bb(B) ) = b_11 b_22 - b_21 b_12

And the product is:

det( bb(A) )det( bb(B) ) = (a_11 a_22 - a_21 a_12)(b_11 b_22 - b_21 b_12)

" " = a_11 a_22b_11 b_22 - a_11 a_22b_21 b_12 -
" " a_21 a_12b_11 b_22 + a_21 a_12b_21 b_12

And if we look at the matrix product, we have:

bb(AB) = ( (a_11b_11+a_12b_21,a_11b_12+a_12b_22), (a_21b_11+a_22b_21, a_21b_12+a_22b_22) )

Leading to:

det(bb(AB)) = (a_11b_11+a_12b_21)(a_21b_12+a_22b_22) -
" " ( a_11b_12+a_12b_22) (a_21b_11+a_22b_21)

" " = a_11b_11a_21b_12 + a_12b_21a_21b_12 +
" " a_11b_11a_22b_22 + a_12b_21a_22b_22 -
" " a_11b_12a_21b_11 - a_12b_22a_21b_11 -
" " a_11b_12a_22b_21 - a_12b_22a_22b_21

" " = a_11b_11a_22b_22 +a_12b_21a_21b_12-
" " a_11b_12a_22b_21 -a_12b_22a_21b_11

" " det( bb(A) )det( bb(B) )

In the general nxxn case the same technique can be used but it is easier to use sums of columns and row vectors along with their inner products.