Sei sulla pagina 1di 5

5.5 SIMILARY The matrix of a linear operator T : V V depends on the basis selected for V.

one of fundamental problems of linear algebra is to choose a basis for V that makes the matrix for V as simple as possible. Often this problem attacked by first finding a matrix for V relative to some simple basis like a standard basis. Usually, this choice does not yield the simplest matrix for V, so that one then looks for away to change the basis in order to simplify the matrix. In order to address a problem like this, we must know how a change of basis affects the matrix of a linear operator. The following theorem explains this.

Theorem 8. Let T : V V be a linear operator on a finite-dimensional vector space. If A is the matrix of T whit respect to a basis B, and A is the matrix of T with respect to a basis B then A = P-1 AP Where P is the transition matrix from B to B. (5.20)

Proof: To establish this theorem, it is convenient to describe the relationship Au = v pictorially by writing Since A is the matrix of T with respect to B, and A is the matrix with respect to B, the following relationship hold for all x in V. [ ] And [ ] These can be written [ ] [ And ] (5.21) [ ] [ ]

[ ] [

To see how the matrices A and A are related, let P be the transition matrix from the B basis to the B basis, so that P-1 is the transition matrix from B to B. Thus, [ ] And [ Which can be written as [ ] [ ] And [ ] [ ] (5.22) ] [ ] [ ]

For compactness, relationship (5.21) and (5.22) can be linked together in a single figure as follows: [ ] P [ ] [ [ ] P-1 ] ]

This figure illustrates that are two ways to obtain the matrix [ matrix [ ] . We can take the bottom path across the figure: that is, [ ] [ ]

from the

(5.23)

Or we can go up the left side, across the top, and down the right side; that is, [ ] [ ]

(5.24)

It follows from (5.23) and (5.24) that [ ] [ ]

(5.25)

For all x in V. it follows from (5.25) and part (b) of exercise 11 that P-1AP = A This proves Theorem 8.

Warning. When applying theorem 8, it is easy to forget weather is the transition matrix from B to B (incorrect) or from B to B (correct). It may help to call B the old basis, B the new basis, A the old matrix, and A the new matrix. Since P is the transition matrix from B to B, P-1 is the transition matrix from B to B. thus, (5.20) can be expressed as: New matrix = P-1(old matrix) P Where P is the transition matrix from the new basis to the old basis.

For a more notational way to remember this formula, we can use the notation [T] B and [T] B introduced in the last section. With this notation (5.20) can be expressed as: [T]B = P-1[T]BP B to B B to B

Example 37 Let T : R2 R2 be defined by (* +) * +

Find the standard matrix for T, that is, the matrix for T relative to the basis B = (e1, e2), where * + * +

And then use theorem 8 to transform this matrix into the matrix for T relative to the basis B = {u1, u2}, where * + * +

Solution: In part (a) of example 33 we found the matrix for T relative to the standard basis to be [ ] * +

Next we need the transition matrix from B to B. For this transition matrix we need the coordinate matrices for the B basis vector relative to the basis B, By inspection, u1 = e1 + e2 u2 = e1 + 2e2 So that, [ ] * + [ ] * +

Thus, the transition matrix from B to B is * The reader can check that * + +

So that by theorem 8 the matrix of T relative to the basis B is

[ ]

+*

+*

This agrees with the result obtained in part (b) of example 33.

This example illustrates that the standard basis for a vector spaces does not necessarily produce the simplest matrix for a linear operator; we saw that the standard matrix * Was not as simple in structure as the matrix * + (5.26) +

relative to the basis B. Matrix (5.26) is an example of a diagonal matrix, that is, a square matrix all of whose non diagonal entries are zeros. Diagonal matrices have many desirable properties. For example, the kth power of a diagonal matrix [ ]

Is

Thus, to raise a diagonal matrix to the kth power, we need only raise each diagonal entry to the kth power. For a non diagonal matrix there is much more computation involved in obtaining the kth power. Diagonal matrices also have other useful properties. In the next chapter we shall discuss the problem of finding bases that produce diagonal matrices for linear operators. Theorem 8 motivates the following definition.

DEFINITION If A and B are square matrices, we say that B is similar to A if there is an invertible matrix P such that B = P-1AP. Note that the equation B = P-1AP can be rewritten as A = PB-1P Letting Q = P-1 yields A = Q-1BQ Which says that A is similar to B. therefore, B is similar to A if and only if A is similar to B; consequently, we shall usually say simply that A and B are similar. In this terminology, Theorem 8 asserts that two matrices representing the same linear operator T : V V with respect to different bases are similar. or A = (P-1)-1BP-1

Potrebbero piacerti anche