Scribd
Upload
6 views

notesU05

Chapter 3 discusses vector spaces and subspaces, defining vector spaces as sets where addition and scalar multiplication are well-defined. It provides examples of real vector spaces, including the set of real numbers, Rn, and matrices, and explains the criteria for a subset to be a subspace. The chapter also introduces the concept of the column space of a matrix as a subspace and its relevance in solving linear equations.

Uploaded by

sevaramchouhan555
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
6 views

notesU05

Chapter 3 discusses vector spaces and subspaces, defining vector spaces as sets where addition and scalar multiplication are well-defined. It provides examples of real vector spaces, including the set of real numbers, Rn, and matrices, and explains the criteria for a subset to be a subspace. The chapter also introduces the concept of the column space of a matrix as a subspace and its relevance in solving linear equations.

Uploaded by

sevaramchouhan555
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Chapter 3: Vector Spaces and Subspaces

3.1. Spaces of Vectors


Now we want to generalize the concept of vector space. In its most general form, we should
begin with the scalars we are allowed to multiply by. They could from any system within which
you can add, subtract, multiply and (except by 0) divide, and all the usual rules of arithmetic
hold. Well, maybe not “all”; we don’t need anything about order (“greater than 0”) or even the
idea that adding 1’s together will never give 0 (there are fields in which that isn’t true). Some
fields are: the rational numbers (integers — whole numbers — over integers), the real numbers,
the complex numbers (things like 3 + 2i where i2 = −1, which you probably learned about in
high school algebra), and the set that I usually refer to as Z2 , but Dan Swiecki says is called B in
computer science: The elements 0 and 1, with the usual rules except that 1 + 1 = 0. (Remember
that I said that sums of 1’s could be 0 in some fields?) So what are the rules that a field must
follow? I don’t want to list them here, but I may post them on the website. Anyway, the only field
we will always (or at least for a long while) use is the real numbers R.
So if all the scalars are just real numbers, what are the vectors? The answer is, anything in a
context in which adding two things and multiplying something by a scalar make sense, and both of
these operations always give you something else in the set. Again, there are some rules that these
operations must follow (these at least are listed in the book, just before the Section 3.1 problem
set), but for present purposes we can just say the operations act as you would expect. Instead
of dwelling much on these rules, let’s list some vector spaces (“over the reals”) that we will use
frequently:

Example. Some real vector spaces:

• The set of real numbers is a vector space over itself: The sum of any two real numbers is a
real number, and a multiple of a real number by a scalar (also real number) is another real
number. And the rules work (whatever they are).

• Rn , for any positive integer n, is a vector space over R: For example, the sum of two lists of
5 numbers is another list of 5 numbers; and a scalar multiple of a list of 5 numbers is another
list of 5 numbers.

• The text uses the symbol M for the set of 2×2 matrices (with real entries — I would have
preferred something more, maybe M2,2 (R), because there is nothing special about 2×2 ma-
trices; but Strang wrote the book. Anyway, the sum of two such matrices is another such
matrix, and a scalar multiple of such a matrix is also such a matrix.

• The set F [−1, 1] of all functions from the interval [−1, 1] into the reals. (The text uses the
set F of functions on the whole real line. I’ve changed that to make the point that the
domain really doesn’t make much difference.) We add two such functions, say f and g, and
we multiply one, say f , by a scalar, say c, “pointwise”, i.e., at each value of x in [−1, 1], the
values of the resulting functions f + g and cf are given by the rule

(f + g)(x) = f (x) + g(x) and (cf )(x) = c(f (x)) .

This is of course the simplest addition and multiplication possible. Here’s a picture:

1
• The text uses Z for the vector space with only one vector, the zero vector. It doesn’t matter
which space this vector is in, because they are all identical from a vector space point of view:
The sum of the vector with itself is itself, all the scalar multiples of the vector are itself.

What we are really interested in at this point are subsets of vector spaces which are vector
spaces in their own rights, with the addition and scalar multiplication inherited from the larger
space. These are called subspaces of the larger space, and it turns out that, for a subset to be a
subspace, all that has to be true is that

• the sum of two vectors in the subset is also in the subset, and

• all scalar multiples of a vector in the subset is also in the subset.

(Well, technically, you need one more thing: The subset can’t be empty, because you need the zero
vector there.)
So, for example:

• In R3 , the set of vectors of the form (a, b, a + b) is a subspace, because

– (a, b, a+b)+(c, d, c+d) = (a+c, b+d, a+b+c+d), so the sum of two vectors of this form is
still of this form (the third entry is the sum of the first two: a+b+c+d = (a+c)+(b+d));
and
– c(a, b, a + b) = (ca, cb, c(a + b)) = (ca, cb, ca + cb).

• In any vector space V , the one-vector vector space Z and V itself are subspaces.

• If v is a vector in some vector space, then the set of all scalar multiples of v form a subspace.
In fact, the span (the set of all linear combinations) of any set of vectors in a vector space is
a subspace — in fact, it is the smallest subspace that contains those vectors.
 
a b
• The set of all 2×2 matrices of the form is a subspace of M. So is the set of all
b a − 2b
 
a b
symmetric matrices: .
b d

• The set of real-valued functions f on [−1, 1] that satisfy f (1/2) = 0 is a subspace of F [−1, 1].
(There is nothing special about the 1/2 here — it could have been any number in [−1, 1] —
but the 0 that follows it must be a zero; otherwise, the sum of two functions of this kind
wouldn’t be another function of this kind, so it wouldn’t be a subspace.)

In view of one of the examples above, we can write down the first (kind of) subspace in which
we are really interested:

2
Definition. For any m × n matrix A, the column space C(A) of A is the span of its columns, a
subspace of Rm .

Why do we care? Because, for any vector x in Rn , Ax is a combination of the columns of A,


with coefficients the components of x. So Ax = b has a solution exactly when b is in C(A).

At this point, switch to the presentation for Unit 5.

3.2. The Nullspace of A: Solving Ax = 0

You might also like