If A is an matrix, each row of A is an n-tuple of real numbers and hence can be considered as a vector in . The m vectors corresponding to the rows of A will be referred to as the row vectors of A. Similarly, each column of A can be considered as a vector in , and we can associate n column vectors with the matrix A.
Let
The row space of A is the set of all 3-tuples of the form
The column space of A is the set of all vectors of the form
Thus, the row space of A is a two-dimensional subspace of , and the column space of A is .
Two row equivalent matrices have the same row space.
Proof
If B is row equivalent to A, then B can be formed from A by a finite sequence of row operations. Thus, the row vectors of B must be linear combinations of the row vectors of A. Consequently, the row space of B must be a subspace of the row space of A. Since A is row equivalent to B, by the same reasoning, the row space of A is a subspace of the row space of B.
∎
To determine the rank of a matrix, we can reduce the matrix to row echelon form. The nonzero rows of the row echelon matrix will form a basis for the row space.
Let
Reducing A to row echelon form, we obtain the matrix
Clearly, and form a basis for the row space of U. Since U and A are row equivalent, they have the same row space, and hence the rank of A is 2.
The concepts of row space and column space are useful in the study of linear systems. A system can be written in the form
In Chapter 1 we used this representation to characterize when a linear system will be consistent. The result, Theorem 1.3.1, can now be restated in terms of the column space of the matrix.
A linear system is consistent if and only if b is in the column space of A.
If b is replaced by the zero vector, then (1) becomes
It follows from (2) that the system will have only the trivial solution if and only if the column vectors of A are linearly independent.
Let A be an m×n matrix. The linear system is consistent for every if and only if the column vectors of A span . The system has at most one solution for every if and only if the column vectors of A are linearly independent.
Proof
We have seen that the system is consistent if and only if b is in the column space of A. It follows that will be consistent for every if and only if the column vectors of A span . To prove the second statement, note that, if has at most one solution for every b, then in particular the system can have only the trivial solution, and hence the column vectors of A must be linearly independent. Conversely, if the column vectors of A are linearly independent, has only the trivial solution. Now, if and were both solutions of , then − would be a solution of :
It follows that , and hence must equal .
∎
Let A be an matrix. If the column vectors of A span , then n must be greater than or equal to m, since no set of fewer than m vectors could span . If the columns of A are linearly independent, then n must be less than or equal to m, since every set of more than m vectors in is linearly dependent. Thus, if the column vectors of A form a basis for , then n must equal m.
An matrix A is nonsingular if and only if the column vectors of A form a basis for .
In general, the rank and the dimension of the null space always add up to the number of columns of the matrix. The dimension of the null space of a matrix is called the nullity of the matrix.
If A is an matrix, then the rank of A plus the nullity of A equals n.
Proof
Let U be the reduced row echelon form of A. The system is equivalent to the system . If A has rank r, then U will have r nonzero rows, and consequently, the system will involve r lead variables and free variables. The dimension of will equal the number of free variables.
∎
Let
Find a basis for the row space of A and a basis for . Verify that dim .
SOLUTION
The reduced row echelon form of A is given by
Thus, is a basis for the row space of A, and A has rank 2. Since the systems and are equivalent, it follows that x is in if and only if
The lead variables and can be solved for in terms of the free variables and x4:
Let and . It follows that consists of all vectors of the form
The vectors and form a basis for . Note that
The matrices A and U in Example 3 have different column spaces; however, their column vectors satisfy the same dependency relations. For the matrix U, the column vectors and are linearly independent, while
The same relations hold for the columns of A: The vectors and are linearly independent, while
In general, if A is an matrix and U is the row echelon form of A, then, since if and only if , their column vectors satisfy the same dependency relations. We will use this property to prove that the dimension of the column space of A is equal to the dimension of the row space of A.
If A is an matrix, the dimension of the row space of A equals the dimension of the column space of A.
Proof
If A is an matrix of rank r, the row echelon form U of A will have r leading 1’s. The columns of U corresponding to the leading 1’s will be linearly independent. They do not, however, form a basis for the column space of A, since, in general, A and U will have different column spaces. Let denote the matrix obtained from U by deleting all the columns corresponding to the free variables. Delete the same columns from A and denote the new matrix by . The matrices and are row equivalent. Thus, if x is a solution of , then x must also be a solution of . Since the columns of are linearly independent, x must equal 0. It follows from the remarks preceding Theorem 3.6.3 that the columns of are linearly independent. Since has r columns, the dimension of the column space of A is at least r.
We have proved that, for any matrix, the dimension of the column space is greater than or equal to the dimension of the row space. Applying this result to the matrix , we see that
Thus, for any matrix A, the dimension of the row space must equal the dimension of the column space.
∎
We can use the row echelon form U of A to find a basis for the column space of A. We need only determine the columns of U that correspond to the leading 1’s. These same columns of A will be linearly independent and form a basis for the column space of A.
The row echelon form U tells us only which columns of A to use to form a basis. We cannot use the column vectors from U, since, in general, U and A have different column spaces.
Let
The row echelon form of A is given by
The leading 1’s occur in the first, second, and fifth columns. Thus,
form a basis for the column space of A.
Find the dimension of the subspace of spanned by
SOLUTION
The subspace is the same as the column space of the matrix
The row echelon form of X is
The first two columns , of X will form a basis for the column space of X. Thus, dim .
3.144.86.138