Você está na página 1de 28

2.2. Higher-Order Determinants The 1 1 matrix [a] is invertible exactly when a = 0. a b The 2 2 matrix is invertible exactly when c d ad bc = 0.

0. What about a 3 3 matrix? Is there some short of expression which will determine when such a matrix is invertible? The answer is yes (and the answer for larger square matrices is also yes), and it is called the determinant of the matrix, but the formula gets very complicated very fast as the matrix gets bigger; it is easier to work on an individual basis.

Lets proceed one step at a time, starting with the minors of a matrix A. The determinant of the matrix obtained by removing all the entries in the ith row and the jth column of A is called the (i, j)th minor of A and is denoted Mi,j . This assumes you know how to nd the determinant of a smaller matrix!

2 4 3 If A = 4 8 3 , then M1,2 , (1, 2)th minor of A, is 3 6 1 the determinant of 2 4 3 4 8 3 = 4 3 , which is 4 1 3 (3) = 13. 3 1 3 6 1 M2,1 is the determinant of 2 4 3 4 8 3 = 4 3 , which is (4) 1 (6) 3 = 14. 6 1 3 6 1

The (i, j) cofactor of A, denoted Ci,j , is (1)i+j Mi,j . The (1)i+j creates a checkerboard pattern, which changes the signs of some of the minors:

+ + + + +

The (i, j) cofactor of A, denoted Ci,j , is (1)i+j Mi,j . The (1)i+j creates a checkerboard pattern, which changes the signs of some of the minors: Minors + + + + + 26 13 0 14 7 0 36 18 0

The (i, j) cofactor of A, denoted Ci,j , is (1)i+j Mi,j . The (1)i+j creates a checkerboard pattern, which changes the signs of some of the minors: Minors + + + + + 26 13 0 14 7 0 36 18 0 Cofactors 26 13 0 14 7 0 36 18 0

The (i, j) cofactor of A, denoted Ci,j , is (1)i+j Mi,j . The (1)i+j creates a checkerboard pattern, which changes the signs of some of the minors: Minors
+ + + + +

Cofactors 13 0 14 7 0 +36 18 0

13 0 14 7 0 +36 18 0

26

26

The (i, j) cofactor of A, denoted Ci,j , is (1)i+j Mi,j . The (1)i+j creates a checkerboard pattern, which changes the signs of some of the minors: Minors Cofactors 26 13 0 14 7 0 36 +18 0

+ + + +
+

26 +13 0 +14 7 0 36 18 0

Now (at last) we can nd the determinant of A, which is dened to be A1,1 C1,1 + A1,2 C1,2 + A1,3 C1,3 + + A1,n C1,n . The determinant of a matrix A is denoted by |A| or det A. Here, the determinant is (2)(26) + (4)(13) + (3)(0) = 0. This method is called Expansion by Minors.

A more interesting example, with some of the details omitted, is: 2 1 3 7 3 7 3 3 2 =2+ 7 9

2 1 +7 9 3

2 1 +3+ 9 3

3 7

= 2 (3 9 2 7) 7 (1 9 2 3) + 3 (1 7 3 3) = 1

As you can see, calculating the determinant of a 3 3 matrix requires calculating the determinant of 3 2 2 matrices. Calculating the determinant of a 4 4 matrix requires calculating the determinant of 4 3 3 matrices, each of which requires the determinants of 3 22 matrices. This makes a total of 12 2 2 matrices. Calculating the determinant of a 5 5 matrix requires calculating the determinant of 5 4 4 matrices, which will require the determinant of 60 2 2 matrices.

In general, the determinant of an n n matrix 1 requires the determinants of n! 2 2 matrices. This 2 is not an ecient procedure! (n! = 1 2 . . . n. 10! is around 3 million, 70! is bigger than a googol.) So how can we cut down the computation time?

First of all, we can expand along any row and get the same answer, not just the rst one. Also, we can expand along any column. So if some row or column has a lot of 0s in it, we can cut down the number of computations. We need to obey the checkerboard pattern, so the rst determinant might be subtracted instead of being added.

3 4 0 6

2 17 0 1

8 1 4 7

0 2 0 0

3 4 0 6

2 17 0 1

8 1 4 7

0 2 0 0

3 4 0 6

2 17 0 1

8 1 4 7

0 4 2 =0 0 0 6 0 +0 + + + + + + 3 4 6
+ +

17 0 1 2 17 1

1 3 4 +2+ 0 7 6 8 3 1 + 0 + 4 7 0

2 0 1

8 4 7 2 17 0 8 1 4

Remember:

3 4 0 6

2 17 0 1

8 1 4 7

0 3 2 =2 0 0 6 0

2 0 1

8 4 7

3 4 0 6

2 17 0 1

8 1 4 7

0 3 2 =2 0 0 6 0

2
0

8
4

3 4 0 6

2 17 0 1

8 1 4 7

0 3 2 =2 0 0 6 0 =2

2 0 1

8 4 7 2 1 8 3 +0+ 7 6 +4 3 6 2 1 8 7

+ Remember:

+ +

3 4 0 6

2 17 0 1

8 1 4 7

0 3 2 =2 0 0 6 0 = 2 4

2 0 1 3 6

8 4 7 2 1

3 4 0 6

2 17 0 1

8 1 4 7

0 3 2 =2 0 0 6 0 = 2 4

2 0 1 3 6

8 4 7 2 1

= 2 4 (3 1 2 6) = 72 Only one 2 2 determinant had to be calculated here!

Now suppose we have an upper triangular matrix* Maybe we want to nd the determinant of: 1 0 0 0 0 3 2 0 0 0 2 1 3 0 0 4 2 2 5 0 3 2 1 8

* An upper triangular matrix is a matrix A where Ai,j = 0 whenever j < i.

1 0 0 0 0

3 2 0 0 0

2 1 3 0 0

4 2 2 5 0

3 2 =1 1 8

2 0 0 0

1 3 0 0

2 2 5 0

3 2 1 8

1 0 0 0 0

3 2 0 0 0

2 1 3 0 0

4 2 2 5 0

3 3 2 =12 0 1 0 8

2 5 0

2 1 8

1 0 0 0 0

3 2 0 0 0

2 1 3 0 0

4 2 2 5 0

3 5 2 =123 0 1 8

1 8

1 0 0 0 0

3 2 0 0 0

2 1 3 0 0

4 2 2 5 0

3 2 = 1 2 3 5 |8| 1 8

1 0 0 0 0

3 2 0 0 0

2 1 3 0 0

4 2 2 5 0

3 2 = 12358 1 8

Where do these numbers come from?

3
2

0 0 0 0

2 1
3

0 0 0

4 2 2
5

0 0

3 2 = 12358 1
8

The determinant of an upper (or lower) triangular matrix is the product of the entries on the diagonal.

We could use this fact, if we knew how to get a matrix into an upper triangular form . . .

Você também pode gostar