Definition.
Upper triangular matrix
An upper triangular matrix is a square matrix whose elements below the diagonal are all zeros, that is:
Note that any of the entries (e.g. , ) can be zero as well.
Example.
Upper triangular matrices
The following matrices are all examples of upper triangular matrices:
Definition.
Lower triangular matrix
A lower triangular matrix is a square matrix whose elements above the diagonal are all zeros, that is:
Again, any of the entries (e.g. , ) can be zero as well.
Example.
Lower triangular matrices
Below are some examples of lower triangular matrices:
Definition.
Triangular matrices
A matrix that is either an upper or lower triangular matrix is called a triangular matrix.
Theorem.
Transpose of a triangular matrix
The following statements are true:
If is an upper triangular matrix, then is a lower triangular matrix.
If is a lower triangular matrix, then is an upper triangular matrix.
Moreover, the diagonal entries of a triangular matrix will remain the same after taking its transpose.
Proof. Let be an upper triangular matrix:
We now take the transpose of to get:
Notice that is a lower triangular matrix. Now, consider the case when is a lower triangular matrix of the form . Taking the transpose of gives us , which means that is an upper triangular matrix.
Finally, since triangular matrices are square by definitionlink, taking their transpose does not change the diagonal entries by propertylink. This completes the proof.
■
Theorem.
Determinant of an upper triangular matrix
The determinant of an upper triangular matrix is equal to the product of its diagonal entries, that is:
Proof. We will prove this by induction. Consider the base case in which our upper triangular matrix is like so:
The determinant of is:
Therefore, the theorem holds for the base case. Now, we assume that the theorem holds for the case. Our goal is to show that the theorem holds for the upper triangular matrix:
We compute the determinant of by cofactor expansionlink along the first column:
Notice how we are finding the determinant of an matrix on the right-hand side. We now use the inductive assumption that the theorem holds for the case:
By the principle of mathematical induction, the theorem therefore holds for the general case. This completes the proof.
■
Theorem.
Determinant of a lower triangular matrix
The determinant of a lower triangular matrix is equal to the product of its diagonal entries, that is:
Proof. Let be a lower triangular matrix. with diagonal entries , , , . By theoremlink, the transpose of is an upper triangular matrix. Now, by theoremlink, we know that the determinant of an upper triangular matrix is equal to the product of its diagonal entries. Because is a square matrix, the diagonal entries of and are the same. This means that is:
Next, by theoremlink, we have that and so:
This means that the determinant of a lower triangular matrix is also equal to the product of its diagonal entries! This completes the proof.
■
Theorem.
Determinant of a triangular matrix is equal to the determinant of its transpose
If is a triangular matrix, then:
Proof. If is a triangular matrix, then its determinant is equal to the product of its diagonal entries by theoremlink and theoremlink. Taking the transpose of a square matrix does not change its diagonal entries by theoremlink, which means that the diagonal entries of are the same as those of . Therefore, the determinant of is equal to the determinant of . This completes the proof.
■
Theorem.
Product of two triangular matrices is also a triangular matrix
The following statements are true:
Proof. We will prove the first statement and then use it to prove the second statement. Consider the following two upper triangular matrices:
The product is:
To show that is an upper triangular matrix, we must show that the -th entry of the -th row of , denoted by , is equal to zero when . From basic matrix multiplication, is:
Now, consider the case when . By definition, we have that:
since is an upper triangular matrix, for .
since is an upper triangular matrix, for .
Let's now compute the summation:
Therefore, is an upper triangular matrix.
Now, let's prove the second statement that the product of two lower triangular matrices is also a lower triangular matrix. Suppose matrices and are lower triangular matrices. By theoremlink, we have that:
By theoremlink, the transpose of a lower triangular matrix is an upper triangular matrix. Therefore, and are both upper triangular matrices. We have just proven that the product of two upper triangular matrices is an upper triangular matrix and thus is an upper triangular matrix.
We now take the transpose of both sides of to get:
Since is an upper triangular matrix, its transpose is a lower triangular matrix by theoremlink. This proves that is a lower triangular matrix. This completes the proof.
NOTE
We could have explicitly computed in to get:
Clearly, we can observe that is an upper triangular matrix. However, this proof is not rigorous because it is based merely on observation.
■
Theorem.
Triangular matrix is invertible if and only if every diagonal entry is non-zero
Let be a triangular matrix. is invertible if and only if every diagonal entry of is non-zero.
Proof. We first prove the forward proposition. Assume is an invertible upper triangular matrix. We will prove the case for a lower triangular matrix later.
Consider the following homogeneous system:
We can express this in matrix form like so:
By theoremlink, since is invertible, has only the trivial solution of .
Now, let's assume one of the diagonals is while the other diagonal entries are all non-zero. The -th row of the system can be written as the following linear equation:
We can solve for , , , using the linear equations below row . For instance, to solve for , we focus on the last linear equation. Since can be any number except zero, the following holds:
Substituting into the second-last linear equation of gives:
We repeat this process until we get . Therefore, becomes:
However, because we have assumed that is zero, can be any value and is no longer restricted to . This contradicts the fact that in must be the zero vector. Therefore, we conclude that if is an invertible upper triangular matrix, then every diagonal entry of must be zero.
* * *
Let's now prove the converse. Assume that is an upper triangular matrix and that every diagonal entry of is non-zero. By theoremlink, the determinant of an upper triangular matrix is equal to the product of its diagonal entries:
Since every diagonal entry of is non-zero, we have that . By theoremlink, we conclude that is invertible. We have now managed to prove the theorem for upper triangular matrices.
* * *
Let's now prove that the theorem also holds for lower triangular matrices. Let be an lower triangular matrix. Assume is invertible. By theoremlink, is an upper triangular matrix. Because is invertible, is also invertible by theoremlink. We have just proven that if is an upper triangular matrix, then every diagonal entry of is non-zero. Because taking the transpose of a square matrix does not change its diagonal entries, we conclude that every diagonal entry of is non-zero.
We now prove the converse. Let be an lower triangular matrix. Assume every diagonal entry of is non-zero. By theoremlink, the determinant of is:
Since every diagonal entry of is non-zero, we have that . By theoremlink, we conclude that is invertible.
We have now shown that the theorem holds for both upper and lower triangular matrices. This completes the proof.
■
Theorem.
Inverse of a triangular matrix is also triangular with the diagonal entries becoming reciprocals
The following statements are true:
if is an invertible upper triangular matrix, then is also an upper triangular matrix.
if is an invertible lower triangular matrix, then is also a lower triangular matrix.
Specifically, the diagonal entries of the inverse matrix are the reciprocals of the diagonal entries of the original matrix.
Proof. Let's start by proving the first statement. We assume is an invertible upper triangular matrix. Since is invertible, exists. Let's express as a collection of column vectors:
Now, the product is:
Where the second equality holds by theoremlink. By definitionlink of invertibility, we have that . This means that:
Where , , , are the standard basis vectors of . Equating the columns gives:
Let's focus on the first equation . In matrix form, this is:
The corresponding system of linear equations is:
Focus on the last equation. Since any entry can be any value, the only way for the equation to hold is if . Next, we substitute into the second-last equation and conclude that using the same logic. We keep on repeating this to conclude that:
Now, we substitute into the first equation of to get:
Next, we focus on . The matrix form is:
The corresponding linear system is:
Again, we perform the same steps to conclude:
Substitute this into the second equation of to get:
In general, for the system , we have:
The value of a diagonal entry is:
The inverse matrix of defined as is therefore:
This means that is an upper triangular matrix.
* * *
Let's now prove the second statement. Let be an invertible lower triangular matrix. By theoremlink, is an upper triangular matrix. We have just proven that the inverse of an upper triangular matrix is also an upper triangular matrix, that is, is an upper triangular matrix. Now, recall the following theoremlink:
Therefore, is an upper triangular matrix. This means that is a lower triangular matrix by theoremlink. This completes the proof.
■
Theorem.
Diagonal entries of the product of two triangular matrices
Let and be both lower or upper triangular matrices in which:
the diagonal entries of are , , , .
the diagonal entries of are , , , .
The diagonal entries of the product are , , , .
Proof. We will first prove the case for when matrices and are upper triangular matrices. For reference, the product is:
By definition, we have that:
since is an upper triangular matrix, for .
since is an upper triangular matrix, for .
Let's now compute the summation. The diagonal entries of are:
This shows that the theorem holds for upper triangular matrices.
* * *
Let's now prove the case for when and are lower triangular matrices. By theoremlink, we have that:
By theoremlink, and are upper triangular matrices. Since and are square matrices, and have the same diagonals as and respectively. We have just proven that the diagonals of the product of two upper triangular matrices, , is for . We now take the transpose of and conclude that the diagonal entries of are also for .
This completes the proof.
■