1
00:00:05,550 --> 00:00:08,770
In this video we look at the
subjective addition and

2
00:00:08,770 --> 00:00:11,668
subtraction of matrices, and we
also look at scalar

3
00:00:11,668 --> 00:00:14,888
multiplication of matrices. To
do that, we're going to need

4
00:00:14,888 --> 00:00:18,108
some matrices, so here are some
matrices that I've already

5
00:00:18,108 --> 00:00:21,328
prepared, and you'll see we've
got four matrices here, and

6
00:00:21,328 --> 00:00:25,514
they've all got different sizes.
So the first thing we need to do

7
00:00:25,514 --> 00:00:30,022
is just remind ourselves as to
how we look at the size of a

8
00:00:30,022 --> 00:00:34,208
matrix, so we count up the
number of rows and the number of

9
00:00:34,208 --> 00:00:35,818
columns. So this matrix A.

10
00:00:35,900 --> 00:00:41,318
We say is a two by two matrix
because it's got two rows and

11
00:00:41,318 --> 00:00:46,684
two columns. Matrix B's got
three rows and two columns, so

12
00:00:46,684 --> 00:00:49,360
that's a three by two matrix.

13
00:00:50,220 --> 00:00:55,095
And we can clearly see that
matrix C is a two by three

14
00:00:55,095 --> 00:00:59,595
matrix, two rows and three
columns, and matrix D is a three

15
00:00:59,595 --> 00:01:02,970
by two matrix with three rows
and two columns.

16
00:01:03,620 --> 00:01:06,920
Now, when it comes to adding and
subtracting matrices, we can

17
00:01:06,920 --> 00:01:10,820
only do it when the two matrices
have the same size. That is,

18
00:01:10,820 --> 00:01:15,020
when they got this both got the
same number of rows and the same

19
00:01:15,020 --> 00:01:18,320
number of columns and two
matrices that have the same size

20
00:01:18,320 --> 00:01:21,620
are said to be compatible, and
when they're compatible we can

21
00:01:21,620 --> 00:01:23,120
add them and subtract them.

22
00:01:23,840 --> 00:01:25,891
So if we return to our four

23
00:01:25,891 --> 00:01:31,733
matrices. We see that of these
four matrices, the only two that

24
00:01:31,733 --> 00:01:37,169
are compatible or matrix B and
matrix D. They have the same

25
00:01:37,169 --> 00:01:43,511
size 3 rows and two columns. So
what that means is that we can

26
00:01:43,511 --> 00:01:50,759
find B + D and we can find
B -- D and we can find D

27
00:01:50,759 --> 00:01:56,648
-- B. So we can add and subtract
matrices, B&D because they have

28
00:01:56,648 --> 00:01:58,007
the same size.

29
00:01:58,120 --> 00:02:03,240
Because they're compatible, we
can't add A&B because they have

30
00:02:03,240 --> 00:02:08,360
different sizes. We can't add
C&A because they have different

31
00:02:08,360 --> 00:02:13,130
sizes. Might be worth noting
that where we define the matrix

32
00:02:13,130 --> 00:02:18,594
C transpose. C transpose. That's
where the rows become columns

33
00:02:18,594 --> 00:02:23,582
would have. Two columns, because
each row would turn into a

34
00:02:23,582 --> 00:02:27,998
column, it would have three
rows, so C transpose would be a

35
00:02:27,998 --> 00:02:31,678
three by two matrix. So C
transpose is also compatible

36
00:02:31,678 --> 00:02:36,830
with B&D. So we can add C
transpose to be or today, but we

37
00:02:36,830 --> 00:02:39,406
can't add C to be or today.

38
00:02:40,970 --> 00:02:44,402
Once we found two matrices that
are compatible that these two

39
00:02:44,402 --> 00:02:48,458
matrices that have got the same
size, then we need to know how

40
00:02:48,458 --> 00:02:52,826
to actually add them up. So we
look at our matrices B&D and see

41
00:02:52,826 --> 00:02:56,570
how we go through this process.
So here's our Matrix B and

42
00:02:56,570 --> 00:03:00,314
here's our Matrix D and I've
written them with a plus sign

43
00:03:00,314 --> 00:03:03,746
between them and underneath I've
written them out again with a

44
00:03:03,746 --> 00:03:09,674
minus sign. So this is B + D and
this is B -- D. So how do we do

45
00:03:09,674 --> 00:03:11,234
the addition? Well, it's quite

46
00:03:11,234 --> 00:03:15,280
straightforward. All we do
is we were adding we add the

47
00:03:15,280 --> 00:03:18,660
elements that are in the
same position. We call that

48
00:03:18,660 --> 00:03:19,674
the corresponding position.

49
00:03:20,960 --> 00:03:25,342
So because the five is in the
first row on the 1st column and

50
00:03:25,342 --> 00:03:29,724
the two is in the first row and
the first column, they get added

51
00:03:29,724 --> 00:03:34,732
together. So we do 5 + 2 which
is 7 and that gives us the entry

52
00:03:34,732 --> 00:03:38,175
in the first row and the first
column of our answer.

53
00:03:39,060 --> 00:03:44,086
We do the same with all the
elements, so the minus one is in

54
00:03:44,086 --> 00:03:48,753
the 2nd row and the first
column. So we add that to the

55
00:03:48,753 --> 00:03:54,138
zero in the 2nd row and the
first column. So we do minus 1 +

56
00:03:54,138 --> 00:03:59,164
0, which gives us minus one and
we can continue to do that for

57
00:03:59,164 --> 00:04:04,190
all six elements of the matrix.
So 1 + 4 because the one and

58
00:04:04,190 --> 00:04:08,498
four are in corresponding
positions gives us 5 -- 2 + --

59
00:04:08,498 --> 00:04:13,360
2. Gives us minus four and that
goes up here because it's in the

60
00:04:13,360 --> 00:04:17,320
first row and the second column
first row on the second column

61
00:04:17,320 --> 00:04:21,940
for three and the one get added
to give us four and the nought

62
00:04:21,940 --> 00:04:25,570
and the minus one get added to
give us minus one.

63
00:04:26,810 --> 00:04:31,282
And so that's how we do matrix
addition. So just to recap, we

64
00:04:31,282 --> 00:04:35,754
have to have two matrices that
have the same size and then when

65
00:04:35,754 --> 00:04:40,570
we have two matrices at the same
size we add them by adding the

66
00:04:40,570 --> 00:04:43,322
elements that are in
corresponding positions. And so

67
00:04:43,322 --> 00:04:47,794
the answer we get is the same
size as the two matrices that

68
00:04:47,794 --> 00:04:48,826
we've added together.

69
00:04:50,120 --> 00:04:53,200
Now the principles of
subtraction are exactly the

70
00:04:53,200 --> 00:04:57,050
same. We deal with elements that
are in the corresponding

71
00:04:57,050 --> 00:05:00,900
positions, but obviously this
time we subtract rather than add

72
00:05:00,900 --> 00:05:07,830
them. So we do 5 -- 2 to get
three, we do minus 1 -- 0 to get

73
00:05:07,830 --> 00:05:14,186
minus one. We do 1 -- 4 to
get minus three. That's done the

74
00:05:14,186 --> 00:05:18,729
elements in the first column
with the elements in the second

75
00:05:18,729 --> 00:05:24,511
column minus 2 -- -- 2 becomes
minus 2 + 2, which is 0.

76
00:05:25,340 --> 00:05:31,268
3 -- 1 gives us 2
and 0 -- -- 1 is 0

77
00:05:31,268 --> 00:05:33,548
+ 1 which is 1.

78
00:05:34,570 --> 00:05:39,666
And there's our answer. So
when we do B -- D, This is the

79
00:05:39,666 --> 00:05:44,034
answer. Again, a matrix of the
same size as B&D, so that

80
00:05:44,034 --> 00:05:47,310
illustrates how we do matrix
addition and subtraction. We

81
00:05:47,310 --> 00:05:51,314
have to have two matrices
which have the same size in

82
00:05:51,314 --> 00:05:56,046
order to be compatible. And
then what we do is we add or

83
00:05:56,046 --> 00:06:00,050
subtract the elements that are
in the same positions. We call

84
00:06:00,050 --> 00:06:00,778
corresponding elements.

85
00:06:01,910 --> 00:06:05,430
Now Matrix obviously has the
same size itself, so we can

86
00:06:05,430 --> 00:06:09,590
always add a matrix to itself,
and we're going to do that now

87
00:06:09,590 --> 00:06:13,430
with the Matrix A. So we're
going to add matrix A to

88
00:06:13,430 --> 00:06:17,590
itself, so into a plus a. So
here's Matrix A and what adding

89
00:06:17,590 --> 00:06:21,110
matrix a onto it. And because
it's the same matrix, clearly

90
00:06:21,110 --> 00:06:25,270
it's not the same size that
both 2 by two matrices, so we

91
00:06:25,270 --> 00:06:26,870
go through the standard
procedure.

92
00:06:28,690 --> 00:06:31,910
When we add elements that are
in corresponding positions, so

93
00:06:31,910 --> 00:06:35,774
the four gets added to the
four, which gives us 8, the

94
00:06:35,774 --> 00:06:39,960
three gets added to the three
to give us 6 not getting to

95
00:06:39,960 --> 00:06:43,502
nought, which gives us nought
and minus one gets added to

96
00:06:43,502 --> 00:06:45,756
minus one. To give this minus
2.

97
00:06:47,240 --> 00:06:53,274
So matrix a + A is this matrix
here with entries 860 and minus

98
00:06:53,274 --> 00:06:59,739
two and we used to writing A
plus a in a shorthand form as a

99
00:06:59,739 --> 00:07:01,894
+ A = 2 A.

100
00:07:02,720 --> 00:07:06,260
One lot of a there's another lot
of a gives us two lots of a.

101
00:07:06,960 --> 00:07:11,302
So this matrix that we found
here, we can refer to as 2A.

102
00:07:12,110 --> 00:07:16,816
And if we look at the entries in
this matrix and compare them

103
00:07:16,816 --> 00:07:21,884
with the entries of a, we see
that each of the entries is just

104
00:07:21,884 --> 00:07:28,038
twice the entries of a 2 * 4 is
eight 2 * 3 or 6 two times

105
00:07:28,038 --> 00:07:32,744
naughties nought 2 * -- 1 is
minus two, and so this process

106
00:07:32,744 --> 00:07:36,726
illustrates how we do we call
scalar multiplication. We take a

107
00:07:36,726 --> 00:07:41,432
matrix and we multiply it by a
number. All that happens is that

108
00:07:41,432 --> 00:07:43,242
every element inside the matrix.

109
00:07:43,360 --> 00:07:47,596
Gets multiplied by the number,
so in this case the number was

110
00:07:47,596 --> 00:07:51,832
two and we'll do some examples
now, but we use a different

111
00:07:51,832 --> 00:07:56,120
number. So we've seen how we can
do scalar multiplication by

112
00:07:56,120 --> 00:07:59,320
simply multiplying every element
inside our matrix by the number.

113
00:07:59,320 --> 00:08:03,160
The scalar that we're trying to
multiply by. So we'll do a

114
00:08:03,160 --> 00:08:06,680
couple more examples now, so
we're going to workout is going

115
00:08:06,680 --> 00:08:11,160
to five times the matrix B and
I'm going to do 1/2 times the

116
00:08:11,160 --> 00:08:16,830
matrix D. So all I've done is
I've written down what matrix B

117
00:08:16,830 --> 00:08:21,270
is. I'm going to do five times
this matrix. So remember the

118
00:08:21,270 --> 00:08:24,970
rule for scalar multiplication
is the scalar, the number that

119
00:08:24,970 --> 00:08:28,300
we're trying to multiply by
multiplies every entry inside

120
00:08:28,300 --> 00:08:35,662
the matrix. So we get 5
* 5 is 20 five 5 * --

121
00:08:35,662 --> 00:08:39,470
1 is minus five. 5 * 1 is

122
00:08:39,470 --> 00:08:47,156
5. 5 * -- 2 is minus
ten. 5 * 3 is 15 and 5

123
00:08:47,156 --> 00:08:49,052
* 0 is 0.

124
00:08:50,080 --> 00:08:54,448
So this is our answer. This is
the matrix 5B or scalar five

125
00:08:54,448 --> 00:08:58,816
times matrix B. Notice that the
matrix we get to that answer has

126
00:08:58,816 --> 00:09:03,184
the same size, the same order as
the matrix we started from, and

127
00:09:03,184 --> 00:09:07,552
that's fairly obvious. That must
be the case because all we do is

128
00:09:07,552 --> 00:09:10,912
we multiply every element inside
the matrix by the scalar

129
00:09:10,912 --> 00:09:14,944
outside. So we are not creating
any new entries in the Matrix

130
00:09:14,944 --> 00:09:19,648
and we're not losing any. So the
matrix that we get must have the

131
00:09:19,648 --> 00:09:22,020
same size. So when we
started with.

132
00:09:23,540 --> 00:09:27,049
Here's another example. We don't
have to just multiply by whole

133
00:09:27,049 --> 00:09:31,196
numbers or two previous
examples. We did 2 * A and 5 *

134
00:09:31,196 --> 00:09:35,343
B, but we can multiply by any
number, and in this case I'm

135
00:09:35,343 --> 00:09:38,533
choosing to multiply the
fraction or half. So I'm going

136
00:09:38,533 --> 00:09:40,447
to do half times matrix D.

137
00:09:41,130 --> 00:09:45,771
So here it is written out.
Here's matrix D. We do 1/2 times

138
00:09:45,771 --> 00:09:50,412
that number. All we have to do
is do 1/2 times every element

139
00:09:50,412 --> 00:09:51,483
inside the matrix.

140
00:09:52,550 --> 00:09:59,206
So we do 1/2 * 2, which gives us
one 1/2 * 0, which gives us

141
00:09:59,206 --> 00:10:04,690
nought 1/2. Times 4, which gives
us two. That's not the first

142
00:10:04,690 --> 00:10:12,002
column and 1/2 * -- 2, giving us
minus 1/2 * 1, giving us a half

143
00:10:12,002 --> 00:10:15,658
and 1/2 * -- 1, giving us minus

144
00:10:15,658 --> 00:10:20,510
1/2. And so here's our product
matrix and not surprisingly, it

145
00:10:20,510 --> 00:10:23,768
ended up with some fractions.
Then, because we were

146
00:10:23,768 --> 00:10:26,664
multiplying by a fraction to
start off with.

147
00:10:27,830 --> 00:10:32,501
That concludes the video on
addition and subtraction of

148
00:10:32,501 --> 00:10:35,096
matrices and on scalar
multiplication.