Return to Video

https:/.../matrixinversionof2x2matrixf61mb-aspect.mp4

  • 0:05 - 0:09
    Once you know how to multiply 2
    matrices together, a natural
  • 0:09 - 0:13
    question to ask is whether or
    not one matrix can be divided by
  • 0:13 - 0:17
    another matrix, and the answer
    to that question is no, there is
  • 0:17 - 0:19
    no such thing as matrix
  • 0:19 - 0:22
    division. Nevertheless, there's
    another operation that we can
  • 0:22 - 0:25
    introduce which plays a very
    similar role to division, and
  • 0:25 - 0:29
    it's called finding the inverse
    of a matrix. So in this video
  • 0:29 - 0:33
    I'm going to explain what we
    mean by the inverse of a matrix.
  • 0:34 - 0:37
    To start with, let's just look
    at these two matrices that I've
  • 0:37 - 0:40
    written down here, and let's
    find the product of them. We
  • 0:40 - 0:42
    multiply these two together.
  • 0:43 - 0:50
    4 * 1 is 4 added to 3 *
    -- 1 is 4 -- 3 which is 1.
  • 0:51 - 0:58
    4 * -- 3 is minus 12, three 4 +
    12 -- 12 + 12 is 0.
  • 1:00 - 1:06
    1 * 1 is 1 and 1 * -- 1 is minus
    one 1 -- 1 is or is not.
  • 1:07 - 1:12
    1 * -- 3 is minus three 1 * 4 is
    four 4 -- 3 is 1.
  • 1:14 - 1:18
    The point I'm trying to make is
    that when we multiply these two
  • 1:18 - 1:21
    matrices together, the answer
    that we get is an identity
  • 1:21 - 1:23
    matrix. That point is important.
  • 1:27 - 1:30
    Suppose we do the
    multiplication in the opposite
  • 1:30 - 1:35
    order, so 1 -- 3 -- 1 four
    multiplied by 4311. So I've
  • 1:35 - 1:37
    changed the order of the
    multiplication there.
  • 1:39 - 1:43
    One's voice 4 -- 3 * 1 is minus
    three. 4 -- 3 is 1.
  • 1:44 - 1:50
    Once three is 3 -- 3 * 1 is
    minus three, 3 -- 3 is nothing.
  • 1:51 - 1:56
    Minus 1 * 4 is minus 44144 and
    the result of adding them
  • 1:56 - 1:57
    together is 0.
  • 1:58 - 2:04
    Minus 1 * 3 is minus 3414, four
    4 -- 3 is 1, so again the result
  • 2:04 - 2:07
    of multiplying these two
    together is an identity matrix.
  • 2:08 - 2:12
    This leads us to the following
    definition for the inverse of a
  • 2:12 - 2:16
    matrix. If you've got one matrix
    and you multiply it by another
  • 2:16 - 2:20
    matrix, and the answer is an
    identity matrix, then the second
  • 2:20 - 2:24
    matrix is the inverse of the
    first matrix and vice versa. The
  • 2:24 - 2:27
    first matrix is the inverse of
    the second matrix. Now suppose
  • 2:27 - 2:31
    we have a matrix A and we
    multiply it by its inverse and
  • 2:31 - 2:35
    we'll denote the inverse by A to
    the power minus one. Now that's
  • 2:35 - 2:39
    not really a power. That does
    not mean 1 / A.
  • 2:39 - 2:42
    This is the notation that
    will use for the inverse of
  • 2:42 - 2:46
    the matrix A, so if you have
    a matrix a multiplied by its
  • 2:46 - 2:47
    inverse.
  • 2:49 - 2:52
    The answer is an identity
    matrix. That's what we mean
  • 2:52 - 2:56
    by an inverse matrix. The
    same works the other way
  • 2:56 - 2:59
    around as well, so if you
    started with an inverse
  • 2:59 - 3:02
    matrix and multiplied it by
    the original matrix, the
  • 3:02 - 3:05
    answer two would be an
    identity matrix, and that's
  • 3:05 - 3:08
    the important definition of
    an identity matrix.
  • 3:09 - 3:13
    If we want to find an inverse of
    the matrix, there's a simple
  • 3:13 - 3:17
    formula that exists when the
    matrix is a two by two matrix.
  • 3:17 - 3:19
    So let's have a look at the
  • 3:19 - 3:26
    formula Now. Suppose we've
    got the matrix ABCD.
  • 3:27 - 3:30
    The formula for the inverse
    is as follows. The inverse,
  • 3:30 - 3:33
    which as I say is denoted by
    A to the minus one.
  • 3:36 - 3:38
    Is given by 1 divided by.
  • 3:39 - 3:44
    The number is 8 * D. Subtract B
    * C rather strange formula, but
  • 3:44 - 3:49
    you'll see how it develops in a
    minute. 1 / a D minus BC
  • 3:49 - 3:53
    multiplied by the matrix that
    you get when you interchange the
  • 3:53 - 3:58
    A and the D round. So the DB
    comes up here and the A goes
  • 3:58 - 4:03
    down there. And we changed the
    sign but leave leave these
  • 4:03 - 4:07
    elements in place. The BNC in
    place, but change their sign to
  • 4:07 - 4:12
    minus BN minus C, and this
    formula that we have here is the
  • 4:12 - 4:15
    formula. Then for the inverse of
    the matrix A.
  • 4:17 - 4:21
    I'm going to apply this
    formula to try and find the
  • 4:21 - 4:25
    inverse of a matrix that I'll
    give you here. Now, supposing
  • 4:25 - 4:27
    the Matrix is a is 3142.
  • 4:30 - 4:34
    Now applying the formula, we'll
    get that the inverse matrix is
  • 4:34 - 4:41
    equal to 1 / 8 * D. That's 3 *
    2, which is 6 -- B * C, which is
  • 4:41 - 4:42
    1 * 4.
  • 4:44 - 4:47
    Multiplied by a matrix which you
    get when you interchange the A
  • 4:47 - 4:50
    and the D. So the A and the D
  • 4:50 - 4:54
    swap round. So swapping the
    three in the two round will
  • 4:54 - 4:55
    get two and three.
  • 4:56 - 4:59
    And changing the sign of the
    other two elements, but leaving
  • 4:59 - 5:00
    them in the same place.
  • 5:02 - 5:04
    So this is the inverse of matrix
  • 5:04 - 5:09
    A. Let's just tidy it
    up a little bit 1 / 6
  • 5:09 - 5:12
    -- 4 is 1 / 2 or 1/2.
  • 5:16 - 5:19
    And we could leave the answer
    like that. Or we can multiply
  • 5:19 - 5:21
    the factor of 1/2 inside,
    multiplying every element inside
  • 5:21 - 5:25
    by the half. But I'm going to
    leave it like that for now.
  • 5:25 - 5:29
    So this is the inverse of matrix
    A. If we want to check whether
  • 5:29 - 5:33
    it's whether it's the right
    answer or not, all we have to do
  • 5:33 - 5:36
    is multiply these two together
    and the answer that we should
  • 5:36 - 5:38
    get should be an identity
    matrix. So let's just check that
  • 5:38 - 5:42
    what I'll do is I'll workout A
    to the minus 1 * A.
  • 5:44 - 5:47
    So here's A to the minus
    one we've just found.
  • 5:49 - 5:53
    And I'm going to multiply it by
    a, which was three 142.
  • 5:56 - 5:58
    Let's leave the half
    outside for now.
  • 6:00 - 6:04
    To do the matrix multiplication,
    here we've got 2, three or six.
  • 6:05 - 6:09
    Added 2 -- 1 * 4 that's 6 -- 4
    which is 2.
  • 6:11 - 6:12
    2 ones or two.
  • 6:12 - 6:16
    Add it to minus 1 * 2. That's 2
    -- 2, which is nothing.
  • 6:18 - 6:19
    Minus 4 * 3.
  • 6:20 - 6:25
    Plus 3 * 4, which is minus 12 +
    12, which is nothing and finally
  • 6:25 - 6:32
    minus 4 * 1, which is minus four
    added to 3, two six 6 -- 4 is 2.
  • 6:33 - 6:38
    And finally, if we multiply each
    element by the factor of 1/2
  • 6:38 - 6:42
    outside, you'll see that we get
    the identity matrix, so that's a
  • 6:42 - 6:45
    verification that this matrix we
    have found here is the inverse
  • 6:45 - 6:47
    of the matrix we started with.
  • 6:48 - 6:51
    For yourself, you could verify
    that we get the same results by
  • 6:51 - 6:54
    multiplying them the opposite
    way around. Finding a time Zeta
  • 6:54 - 6:57
    minus one, and you'll still find
    that you get an identity matrix,
  • 6:57 - 7:01
    and that's how you use the
    formula to find the inverse of a
  • 7:01 - 7:02
    two by two matrix.
  • 7:03 - 7:07
    Now there are certain situations
    where this won't work, and the
  • 7:07 - 7:11
    reason why it won't work is
    obtained by looking at this
  • 7:11 - 7:12
    quantity in the formula here.
  • 7:13 - 7:17
    If it transpires that a D -- B C
  • 7:17 - 7:21
    is 0. Would be trying to divide
    by zero here.
  • 7:22 - 7:24
    And division by zero is never
  • 7:24 - 7:31
    possible. So what this means is,
    if ABC&D are such that a D -- B
  • 7:31 - 7:32
    C is 0.
  • 7:32 - 7:35
    The inverse of this matrix
    will not exist.
  • 7:36 - 7:40
    Now you may remember that the
    quantity AD minus BC we've met
  • 7:40 - 7:43
    before. It's called the
    determinant of this matrix A.
  • 7:44 - 7:47
    So whenever the determinant of
    the matrix A is 0.
  • 7:48 - 7:50
    The inverse won't exist.
  • 7:51 - 7:55
    Equivalently, whenever A is a
    singular matrix, the inverse
  • 7:55 - 8:01
    won't exist. Let's just have a
    quick look at an example of that
  • 8:01 - 8:07
    sort, supposing we had a matrix
    A which was 326 fourth.
  • 8:08 - 8:14
    If we find a D minus BC for this
    matrix, that's 3 * 4, which is
  • 8:14 - 8:19
    12. Subtract 2 * 6 which is 1212
    -- 12 is zero. We see that the
  • 8:19 - 8:21
    determinant of a is 0.
  • 8:23 - 8:27
    This is a singular matrix, and
    so because it's a singular
  • 8:27 - 8:29
    matrix, no inverse matrix will
  • 8:29 - 8:34
    exist. So the point to remember
    is that not all two by two
  • 8:34 - 8:38
    matrices will have an inverse.
    Those that don't have an inverse
  • 8:38 - 8:41
    accord singular matrices. But
    when a matrix does have an
  • 8:41 - 8:45
    inverse, this formula will
    enable you to find it.
Title:
https:/.../matrixinversionof2x2matrixf61mb-aspect.mp4
Video Language:
English
Duration:
08:50

English subtitles

Revisions