You are about to erase your work on this activity. Are you sure you want to do this?
Updated Version Available
There is an updated version of this activity. If you update to the most recent version of this activity, then your current progress on this activity will be erased. Regardless, your record of completion will remain. How would you like to proceed?
Mathematical Expression Editor
Similar Matrices and Their Properties
Let \(A\) and \(B\) be \(n \times n\) matrices. Then the products \(AB\) and \(BA\) are both \(n \times n\) matrices. In most cases the
products \(AB\) and \(BA\) are not equal.
However, for some pairs of \(n \times n\) matrices \(A\) and \(B\), we are able to find an invertible matrix \(P\)
such that \(PB = AP\). This leads to the following definition.
If \(A\) and \(B\) are \(n \times n\) matrices, we say that \(A\) and \(B\) are similar, if \(B = P^{-1}AP\) for some invertible matrix \(P\).
In this case we write \(A \sim B\).
The following theorem shows that similarity (\(\sim \)) satisfies reflexive, symmetric, and
transitive properties.
Similarity is an equivalence relation, i.e. for \(n \times n\) matrices \(A,B,\) and \(C\),
(a)
\(A \sim A\) (reflexive)
(b)
If \(A \sim B\), then \(B \sim A\) (symmetric)
(c)
If \(A \sim B\) and \(B \sim C\), then \(A \sim C\) (transitive)
item:reflexive It is clear that \(A\sim A\) (let \(P=I\)).
item:symmetric If \(A\sim B,\) then for some invertible matrix \(P\),
Any relation satisfying the reflexive, symmetric and transitive properties is called an
equivalence relation. Theorem th:similarityequivalence proves that similarity between matrices is an
equivalence relation. Practice Problem prob:lessthan gives a good example of a relation that is
NOT an equivalence relation.
As we will see later, similar matrices share many properties. Before proceeding to
explore these properties, we pause to introduce a simple matrix function that we will
continue to use throughout the course.
The trace of an \(n \times n\) matrix \(A\), abbreviated \(\mbox {tr} A\), is defined to be the sum of the main diagonal
elements of \(A\). In other words, if \( A = [a_{ij}]\), then
We may also write \(\mbox {tr}(A) =\sum _{i=1}^n a_{ii}\).
It is easy to see that \(\mbox {tr}(A + B) = \mbox {tr} A + \mbox {tr} B\) and that \(\mbox {tr}(cA) = c \mbox { tr}(A)\) holds for all \(n \times n\) matrices \(A\) and \(B\) and all scalars \(c\). The
following fact is more surprising.
Let \(A\) and \(B\) be \(n \times n\) matrices. Then \(\mbox {tr}(AB) = \mbox {tr}(BA)\).
Write \(A = [a_{ij}]\) and \(B = [b_{ij}]\). For each \(i\), the \((i, i)\)-entry \(d_{i}\) of the matrix \(AB\) is given as follows: \(d_{i} = a_{i1}b_{1i} + a_{i2}b_{2i} + \dots + a_{in}b_{ni} = \sum _{j}a_{ij}b_{ji}\). Hence
Similarly we have \(\mbox {tr}(BA) = \sum _{i}\left (\sum _{j}b_{ij}a_{ji}\right )\). Since these two double sums are the same, we have proven the
theorem.
The following theorem lists a number of properties shared by similar matrices.
If \(A\) and \(B\) are \(n\times n\) matrices and \(A\sim B\), then
(a)
\(\det (A) = \det (B)\),
(b)
\(\mbox {rank}(A) = \mbox {rank}(B)\),
(c)
\(\mbox {tr}(A)= \mbox {tr}(B)\),
(d)
\(A\) and \(B\) have the same characteristic equations, and
(e)
\(A\) and \(B\) have the same eigenvalues.
Let \(B = P^{-1}AP\) for some invertible matrix \(P\).
Similarly, for th:properties_similar_rank\(\mbox {rank} B = \mbox {rank}(P^{-1}AP) = \mbox {rank} A\), because multiplication by an invertible matrix cannot change the
rank. To see this, note that any invertible matrix is a product of elementary
matrices. Multiplying by elementary matrices is equivalent to performing
elementary row (column) operations on \(A\), which does not change the row
(column) space, nor the rank. It follows that similar matrices have the same
rank.
Sharing the five properties in Theorem th:properties_similar does not guarantee that two matrices are
similar. The matrices \(A = \begin{bmatrix} 1 & 1 \\ 0 & 1 \end{bmatrix}\) and \(I = \begin{bmatrix} 1 & 0 \\ 0 & 1 \end{bmatrix}\) have the same determinant, rank, trace, characteristic
polynomial, and eigenvalues, but they are not similar because \(P^{-1}IP = I\) for any invertible
matrix \(P\).
Even though the properties in Theorem th:properties_similar cannot be used to show two matrices are
similar, these properties come in handy for showing that two matrices are NOT
similar.
Are the matrices \(A = \begin{bmatrix} 2 & 1 \\ 1 & -1 \end{bmatrix}\) and \(B = \begin{bmatrix} 3 & 0 \\ 1 & -1 \end{bmatrix}\) similar?
A quick check shows us \(\det A = \det B\), and both matrices are seen
to be invertible, so they have the same rank. However, \(\mbox {tr} A = 1\) and \(\mbox {tr} B = 2\), so the matrices are not
similar.
The next theorem shows that similarity is preserved under inverses, transposes, and
powers:
If \(A\) and \(B\) are \(n\times n\) matrices and \(A\sim B\), then
At the beginning of this section we mentioned that similarity of \(n \times n\) matrices is an
equivalence relation.
An equivalence relation is a binary relation \(R\) on elements of a set \(S\) that has the
following properties:
The reflexive property: \(x R x\) for every \(x \in S\)
The symmetric property: If \(x R y\), then \(y R x\) for every \(x,y \in S\)
The transitive property: If \(x R y\) and \(y R z\), then \(x R z\) for every \(x,y,z \in S\)
Let \(S\) be the set of all positive integers. We can show that the relation “less than”
(symbolized by \(<\)) is NOT an equivalence relation on this set. To see this, note
that “less than” is not reflexive, because \(s<s\) is not true for any positive integer
\(s\).
(a)
Is the relation “less than” symmetric? YesNo
(b)
Is the relation “less than” transitive? YesNo
Another relation between matrices we have studied in this course is that two
matrices can be “row equivalent”. Is the relation “row equivalent”
(a)
reflexive? YesNo
(b)
symmetric? YesNo
(c)
transitive? YesNo
Use Theorem 7 to show that \(A\) and \(B\) are not similar. \(A = \begin{bmatrix} 1 & 2 \\ 2 & 1 \end{bmatrix}\), \(B = \begin{bmatrix} 1 & 1\\ -1 & 1 \end{bmatrix}\)
Use Theorem 7 to show that \(A\) and \(B\) are not similar. \(A = \begin{bmatrix} 3 & 1 \\ -1 & 2 \end{bmatrix}\), \(B = \begin{bmatrix} 2 & -1 \\ 3 & 2 \end{bmatrix}\)
Use Theorem 7 to show that \(A\) and \(B\) are not similar. \(A = \begin{bmatrix} 1 & 2 & -3 \\ 1 & -1 & 2 \\ 0 & 3 & -5 \end{bmatrix}\), \(B = \begin{bmatrix} -2 & 1 & 3 \\ 6 & -3 & -9 \\ 0 & 0 & 0 \end{bmatrix}\)
We will refer to the eigenvectors you listed above as \(\vec {x}_1\), \(\vec {x}_2\) and \(\vec {x}_3\).
Form matrix \(P\)
whose columns are the eigenvectors \(\vec {x}_1\), \(\vec {x}_2\), and \(\vec {x}_3\) you found in the previous part. Use
technology to find \(P^{-1}\).
Let \(\lambda \) be an eigenvalue of \(A\) with corresponding eigenvector \(\vec {x}\). If \(B = P^{-1}AP\) is similar to \(A\), show
that \(P^{-1}\vec {x}\) is an eigenvector of \(B\) corresponding to \(\lambda \).