官术网_书友最值得收藏!

Permutations

In the example on solving systems of linear equations, we swapped the positions of rows 2 and 3. This is known as a permutation

When we are doing triangular factorization, we want our pivot values to be along the diagonal of the matrix, but this won't happen every time—in fact, it usually won't. So, instead, what we do is swap the rows so that we get our pivot values where we want them. 

But that is not their only use case. We can also use them to scale individual rows by a scalar value or add rows to or subtract rows from other rows.

Let's start with some of the more basic permutation matrices that we obtain by swapping the rows of the identity matrix. In general, we have n! possible permutation matrices that can be formed from an nxn identity matrix. In this example, we will use a 3×3 matrix and therefore have six permutation matrices, and they are as follows:

  •  This matrix makes no change to the matrix it is applied on.
  •  This matrix swaps rows two and three of the matrix it is applied on.
  •  This matrix swaps rows one and two of the matrix it is applied on.
  •  This matrix shifts rows two and three up one and moves row one to the position of row three of the matrix it is applied on.
  •  This matrix shifts rows one and two down one and moves row three to the row-one position of the matrix it is applied on.
  •  This matrix swaps rows one and three of the matrix it is applied on.

It is important to note that there is a particularly fascinating property of permutation matrices that states that if we have a matrix  and it is invertible, then there exists a permutation matrix that when applied to A will give us the LU factor of A. We can express this like so:

主站蜘蛛池模板: 山丹县| 治县。| 唐河县| 鄯善县| 文成县| 如皋市| 新干县| 阳谷县| 金溪县| 长治县| 乌拉特前旗| 余江县| 宝清县| 稻城县| 彰化市| 申扎县| 武穴市| 墨江| 南充市| 九江市| 永昌县| 郑州市| 宕昌县| 菏泽市| 平凉市| 二手房| 南乐县| 宜君县| 远安县| 邛崃市| 专栏| 洮南市| 柘城县| 增城市| 酉阳| 明水县| 周宁县| 建始县| 德清县| 台安县| 随州市|