|
We need some background knowledge before we can discuss the definition of
the determinant. We want to form a product by choosing n elements where A is an n by n matrix. There will only be one element from each row and one element from each column in this product. For example, if one element of the product is a21, then no other element in this product will be from row 2 or column 1. Let's look at a 3 by 3 matrix for an example.
We know that we will use an element from each column, so, for consistency, we will order the product this way: a_1a_2a_3. We can fill in the blanks with row numbers. If we choose to begin with a31, then we can choose from rows 1 and 2 for the remaining positions. One possible product formed by these rules is a31a12a23. Another possible product is a11a32a23. There are 3!, or 3*2*1 = 6, of these products. For our n by n matrix, there are n! possible products. All 6 possible products for this 3 by 3 matrix are: a11a22a33, a21a32a13, a31a12a23, a31a22a13, a21a12a33, and a11a32a23. Now, we need to determine which sign ( + or -) should be attached to each product. To do this, you need to order the product with the column numbers increasing as we did above and look at the sequences of row numbers. For the product, a11a22a33, we look at the row sequence (1,2,3). We are looking for inversions, or numbers that are out of order. Since 1 comes before 2, 1 comes before 3, and 2 comes before 3 in the sequence, there are no inversions in this sequence. In the sequence (2,3,1), which comes from the product a21a32a13, there are two inversions because 2 is placed before 1 and 3 is placed before 1. There are also two inversions for the sequence from the product a31a12a23. There are three inversions for (3,2,1) and one inversion each for (2,1,3) and (1,3,2). If the number of inversions is even, then the sign attached to the product is positive. If the number of inversions is odd, then the sign attached to the product is negative. Notice that we did not say that the product was positive or negative. We simply are determining whether the product will be added or subtracted. Definition 5.1 The determinant of a square matrix is the sum of all the n possible signed products formed from the matrix using each row and each column only once for each product. The sign to be attached to the product is the same as the one determined by the formula (-1)N when N is the number of inversions as described above.The determinant of the generic 3 by 3 matrix is: a11a22a33 + a21a32a13 + a31a12a23 - a31a22a13 - 0a21a12a33 - a11a32a23. For the matrix,
the determinant is (0*2*6) + (4*3*4) + (1*2*3) - (1*2*4) - (4*2*6) - (0*3*3) = 0 + 48 + 6 - 8 - 48 - 0 = -2 which is the same as we calculated at the beginning of the chapter. Since this definition is cumbersome to follow, we generally do not compute the determinant by the definition, but it is good to know why the short cuts that we learned at the beginning of the chapter are valid. The determinant is also a good example of an abstract idea that has very important practical uses. |
Send comments on material to Tamara Carter
These pages are maintained by Hilena Vargas
(hvargas@rice.edu)
Updated: August 23, 2000
Copyright ©1995 - 2000 Tamara Lynn Anthony