1
00:00:00 --> 00:00:06
OK, here's linear algebra
lecture seven.

2
00:00:06 --> 00:00:15
I've been talking about vector
spaces and specially the null

3
00:00:15 --> 00:00:23
space of a matrix and the column
space of a matrix.

4
00:00:23 --> 00:00:27
What's in those spaces.

5
00:00:27 --> 00:00:34
Now I want to actually describe
them.

6
00:00:34 --> 00:00:39
How do you describe all the
vectors that are in those

7
00:00:39 --> 00:00:40
spaces?

8
00:00:40 --> 00:00:43
How do you compute these
things?

9
00:00:43 --> 00:00:46
So this is the,
turning the idea,

10
00:00:46 --> 00:00:49
the definition,
into an algorithm.

11
00:00:49 --> 00:00:53
What's the algorithm for
solving A x =0?

12
00:00:53 --> 00:00:58
So that's the null space that
I'm interested in.

13
00:00:58 --> 00:01:03
So can I take a particular
matrix A and describe the

14
00:01:03 --> 00:01:06
natural algorithm,
and I'll execute it for that

15
00:01:06 --> 00:01:08
matrix -- here we go.

16
00:01:08 --> 00:01:11
So let me take the matrix as an
example.

17
00:01:11 --> 00:01:16
So we're definitely talking
rectangular matrices in this

18
00:01:16 --> 00:01:16
chapter.

19
00:01:16 --> 00:01:20
So I'll make,
I'll have four columns.

20
00:01:20 --> 00:01:22
And three rows.

21
00:01:22 --> 00:01:28
Two four six eight and three
six eight ten.

22
00:01:28 --> 00:01:28
OK.

23
00:01:28 --> 00:01:34
If I just look at those
columns, and rows,

24
00:01:34 --> 00:01:41
well, I notice right away that
column two is a multiple of

25
00:01:41 --> 00:01:43.09
column one.

26
00:01:43.09 --> 00:01:49
It's in the same direction as
column one.

27
00:01:49 --> 00:01:52.03
It's not independent.

28
00:01:52.03 --> 00:01:56
I'll expect to discover that in
the process.

29
00:01:56 --> 00:02:01
Actually, with rows,
I notice that that row plus

30
00:02:01 --> 00:02:04
this row gives the third row.

31
00:02:04 --> 00:02:07.82
So the third row is not
independent.

32
00:02:07.82 --> 00:02:12
So, all that should come out of
elimination.

33
00:02:12 --> 00:02:17
So now what I --
my algorithm is elimination,

34
00:02:17 --> 00:02:22
but extended now to the
rectangular case,

35
00:02:22 --> 00:02:28
where we have to continue even
if there's zeros in the pivot

36
00:02:28 --> 00:02:30
position, we go on.

37
00:02:30 --> 00:02:35
OK, so let me execute
elimination for that matrix.

38
00:02:35 --> 00:02:41
My goal is always,
while I'm doing elimination --

39
00:02:41 --> 00:02:44
I'm not changing the null
space.

40
00:02:44 --> 00:02:46
That's very important,
right?

41
00:02:46 --> 00:02:50
I'm solving A x equals zero by
elimination, and when I do these

42
00:02:50 --> 00:02:55
operations that you already
know, when I subtract a multiple

43
00:02:55 --> 00:02:59.69
of one equation from another
equation, I'm not changing the

44
00:02:59.69 --> 00:03:00
solutions.

45
00:03:00 --> 00:03:03
So I'm not changing the null
space.

46
00:03:03 --> 00:03:07
Actually, I changing the column
space, as you'll see.

47
00:03:07 --> 00:03:09
So you have to pay attention.

48
00:03:09 --> 00:03:11
What does elimination leave
unchanged?

49
00:03:11 --> 00:03:15.55
And the answer is the solutions
to the system are not changed

50
00:03:15.55 --> 00:03:19.45
because I'm doing the same thing
to -- I'm doing a legitimate

51
00:03:19.45 --> 00:03:21
operations on the equations.

52
00:03:21 --> 00:03:26
Of course, on the right hand
side it's always zero,

53
00:03:26 --> 00:03:29
and I don't plan to write zero
all the time.

54
00:03:29 --> 00:03:33.65
OK, so I'm really just working
on the left side,

55
00:03:33.65 --> 00:03:37
but the right side is,
is keeping up always zeros.

56
00:03:37 --> 00:03:40
OK, so what's elimination?

57
00:03:40 --> 00:03:46
Well, you know where the first
pivot is and you know what to do

58
00:03:46 --> 00:03:46
with it.

59
00:03:46 --> 00:03:50
So can I just take the first
step below here?

60
00:03:50 --> 00:03:53
So that pivot row is fine.

61
00:03:53 --> 00:03:58
I take two times that row away
from this one and I get zero

62
00:03:58 --> 00:03:58
zero.

63
00:03:58 --> 00:04:01
That's signaling a difficulty.

64
00:04:01 --> 00:04:06
Two, two twos away from the six
leaves me with a two.

65
00:04:06 --> 00:04:10
Two twos away from the eight
leaves me with a four.

66
00:04:10 --> 00:04:13
And now three of those away
from here is zero,

67
00:04:13 --> 00:04:17
again another zero,
three twos away from that eight

68
00:04:17 --> 00:04:20
is the two, three twos away from
that ten is a four.

69
00:04:20 --> 00:04:20
OK.

70
00:04:20 --> 00:04:24
That's the first stage of
elimination.

71
00:04:24 --> 00:04:27
I've got the first column
straight.

72
00:04:27 --> 00:04:32
So of course I move on to the
second column.

73
00:04:32 --> 00:04:36
I look in that position,
I see a zero.

74
00:04:36 --> 00:04:41
I look below it,
hoping for a non-zero that I

75
00:04:41 --> 00:04:43
can do a row exchange.

76
00:04:43 --> 00:04:46
But it's zero below.

77
00:04:46 --> 00:04:50
So that's telling me that that
column is -- well,

78
00:04:50 --> 00:04:55
what it's really going to be
telling me is that that column

79
00:04:55 --> 00:04:59
is a combination of the earlier
columns.

80
00:04:59 --> 00:05:03
It's that second column is
dependent on the earlier

81
00:05:03 --> 00:05:04
columns.

82
00:05:04 --> 00:05:06.68
But I don't stop to think here.

83
00:05:06.68 --> 00:05:10
In that column there's nothing
to do.

84
00:05:10 --> 00:05:12
I go on to the next.

85
00:05:12 --> 00:05:14
So here's the next pivot.

86
00:05:14 --> 00:05:19
So there's the first pivot and
there's the second pivot,

87
00:05:19 --> 00:05:23
and I just keep this
elimination going downwards.

88
00:05:23 --> 00:05:29
So, so the next step keeps the
first row, keeps the second row

89
00:05:29 --> 00:05:32
with its pivot,
so I've got my two pivots,

90
00:05:32 --> 00:05:39
and does elimination to clear
out the column below that pivot.

91
00:05:39 --> 00:05:43
So actually you see the
multiplier is one.

92
00:05:43 --> 00:05:49
It subtracts row two from row
three and produces a row of

93
00:05:49 --> 00:05:50
zeros.

94
00:05:50 --> 00:05:50
OK.

95
00:05:50 --> 00:05:54
That I would call that matrix
U, right?

96
00:05:54 --> 00:05:59
That's our upper -- well,
I can't quite say upper

97
00:05:59 --> 00:06:02
triangular.

98
00:06:02 --> 00:06:08
Maybe upper -- I don't know --
upper something.

99
00:06:08 --> 00:06:12
It's in this so-called echelon
form.

100
00:06:12 --> 00:06:17
The word echelon means,
like, staircase form.

101
00:06:17 --> 00:06:25
It's the, the non-zeros come in
that staircase form.

102
00:06:25 --> 00:06:30
If there was another pivot
here, then the staircase would

103
00:06:30 --> 00:06:31
include that.

104
00:06:31 --> 00:06:35
But here's a case where we have
two pivots only.

105
00:06:35 --> 00:06:40
OK, so actually we've already
discovered the most important

106
00:06:40 --> 00:06:42.12
number about this matrix.

107
00:06:42.12 --> 00:06:44
The number of pivots is two.

108
00:06:44 --> 00:06:49
That number we will call the
rank of the matrix.

109
00:06:49 --> 00:06:52
So let me put immediately.

110
00:06:52 --> 00:07:00
The rank of A -- so I'm telling
you what this word rank means in

111
00:07:00 --> 00:07:02
the algorithm.

112
00:07:02 --> 00:07:06
It's equal to the number of
pivots.

113
00:07:06 --> 00:07:09
And in this case,
two.

114
00:07:09 --> 00:07:16
OK, for me that number two is
the crucial number.

115
00:07:16 --> 00:07:21
OK, now I go to -- you remember
I'm always solving A x equals

116
00:07:21 --> 00:07:25
zero, but now I can solve U x
equals zero, right?

117
00:07:25 --> 00:07:28
Same solution,
same null space.

118
00:07:28 --> 00:07:28
OK.

119
00:07:28 --> 00:07:33
So I could stop here -- why
don't I stop here and do the

120
00:07:33 --> 00:07:36.04
back substitution.

121
00:07:36.04 --> 00:07:39
So now I have to ask you,
how do I describe the

122
00:07:39 --> 00:07:40
solutions?

123
00:07:40 --> 00:07:44
There are solutions,
right, to A x equals zero.

124
00:07:44 --> 00:07:45
I knew there would be.

125
00:07:45 --> 00:07:48
I had three equations in four
unknowns.

126
00:07:48 --> 00:07:51
I certainly expected some
solutions.

127
00:07:51 --> 00:07:53
Now I want to see what are
they.

128
00:07:53 --> 00:07:57
OK, here's the critical step.

129
00:07:57 --> 00:08:01
I refer to it up here as
separating out the pivot

130
00:08:01 --> 00:08:06
variables, the pivot columns,
which are these two.

131
00:08:06 --> 00:08:09
Here I have two pivot columns.

132
00:08:09 --> 00:08:14
Those, obviously,
they're the columns with the

133
00:08:14 --> 00:08:15
pivots.

134
00:08:15 --> 00:08:17
So I have two pivot columns.

135
00:08:17 --> 00:08:23
And I have the other columns,
I'll call free.

136
00:08:23 --> 00:08:26
These are free columns,
OK.

137
00:08:26 --> 00:08:29
Why do I use those words?

138
00:08:29 --> 00:08:33
Why do I use that word free?

139
00:08:33 --> 00:08:40
Because now I want to write,
I want to find the solutions to

140
00:08:40 --> 00:08:42
U x equals zero.

141
00:08:42 --> 00:08:44
Here is the way I do it.

142
00:08:44 --> 00:08:53
These free columns I can assign
any number freely to those --

143
00:08:53 --> 00:09:00
to the variables x2 and x4,
the ones that multiply columns

144
00:09:00 --> 00:09:02
two and four.

145
00:09:02 --> 00:09:07
So I can assign anything I like
to x2 and x4.

146
00:09:07 --> 00:09:13
And then I can solve the
equations for x1 and x3.

147
00:09:13 --> 00:09:16
Let me say that again.

148
00:09:16 --> 00:09:21
If I give -- let me,
let me assign.

149
00:09:21 --> 00:09:27
So, so one particular x is to
assign, say, the value one to

150
00:09:27 --> 00:09:30.05
the, to x2, and zero to x4.

151
00:09:30.05 --> 00:09:35
Those are -- that was a free
choice, but it's a convenient

152
00:09:35 --> 00:09:36
choice.

153
00:09:36 --> 00:09:36
OK.

154
00:09:36 --> 00:09:42
Now I want to solve U x equals
zero and find numbers one and

155
00:09:42 --> 00:09:46.45
three, complete the solution.

156
00:09:46.45 --> 00:09:46
OK.

157
00:09:46 --> 00:09:49
Can I write down -- let's see.

158
00:09:49 --> 00:09:55
Shall we just remember what U x
equals zero represents?

159
00:09:55 --> 00:09:57
What are my equations?

160
00:09:57 --> 00:10:03
That first equation is x1 plus
just -- I'm just saying what are

161
00:10:03 --> 00:10:05.71
these matrices meaning.

162
00:10:05.71 --> 00:10:08
That's the first equation.

163
00:10:08 --> 00:10:13
And the second equation was 2x3
+ 4x4=0.

164
00:10:13 --> 00:10:15
Those are my two equations.

165
00:10:15 --> 00:10:16.24
OK.

166
00:10:16.24 --> 00:10:22
Now the point is I can find x1
and x3 by back substitution.

167
00:10:22 --> 00:10:26
So we're building on what we
already know.

168
00:10:26 --> 00:10:31
The new thing is that there
were some free variables that I

169
00:10:31 --> 00:10:35
could give any numbers to.

170
00:10:35 --> 00:10:40.1
And I'm systematically going to
make a choice like this,

171
00:10:40.1 --> 00:10:40
1 and 0.

172
00:10:40 --> 00:10:42
Now what is x3?

173
00:10:42 --> 00:10:45
Let's, let's go backwards,
back up.

174
00:10:45 --> 00:10:49
I look at the last equation.
x3 is zero, from the last

175
00:10:49 --> 00:10:53
equation, because,
because x4 we've set x4 to

176
00:10:53 --> 00:10:57
zero, and then we get x3 as
zero.

177
00:10:57 --> 00:10:57
OK.

178
00:10:57 --> 00:11:00
Now we set x2 to be one,
so what is x1?

179
00:11:00 --> 00:11:02
Negative two,
right?

180
00:11:02 --> 00:11:07
So then I have negative two
plus two, zero and zero,

181
00:11:07 --> 00:11:09
correctly giving zero.

182
00:11:09 --> 00:11:12
There is a vector in the null
space.

183
00:11:12 --> 00:11:16
There is a solution to A x
equals zero.

184
00:11:16 --> 00:11:19
In fact, what solution is that?

185
00:11:19 --> 00:11:23
That simply says that minus two
times the first column plus one

186
00:11:23 --> 00:11:26
times the second column is the
zero column.

187
00:11:26 --> 00:11:28
Of course that's right.

188
00:11:28 --> 00:11:32
Minus two of that column plus
one of that, or minus two of

189
00:11:32 --> 00:11:34
that plus one of that.

190
00:11:34 --> 00:11:38
That solution is --
that, that's just what we saw

191
00:11:38 --> 00:11:43.81
immediately, that the second
column is twice as big as the

192
00:11:43.81 --> 00:11:44
first column.

193
00:11:44 --> 00:11:49
OK, tell me some more vectors
in the null space.

194
00:11:49 --> 00:11:50
I found one.

195
00:11:50 --> 00:11:55
Tell me, how to get a bunch
more immediately out of that

196
00:11:55 --> 00:11:56
one.

197
00:11:56 --> 00:11:58
Just take multiples of it.

198
00:11:58 --> 00:12:01.8
Any multiple of -- I could
multiply this by anything.

199
00:12:01.8 --> 00:12:04
I might as well call it,
I could say,

200
00:12:04 --> 00:12:06
C, some multiple of this.

201
00:12:06 --> 00:12:09
So let me -- so X could be any
multiple of this one.

202
00:12:09 --> 00:12:13
OK, that, that describes now a
line, an infinitely long line in

203
00:12:13 --> 00:12:16.37
four dimensional space.

204
00:12:16.37 --> 00:12:20.14
But -- which is in the null
space.

205
00:12:20.14 --> 00:12:23
Is that the whole null space?

206
00:12:23 --> 00:12:23
No.

207
00:12:23 --> 00:12:27
I've got two free variables
here.

208
00:12:27 --> 00:12:31.11
I made this choice,
one and zero,

209
00:12:31.11 --> 00:12:37
for the free variables,
but I could have made another

210
00:12:37 --> 00:12:39
choice.

211
00:12:39 --> 00:12:42
Let me make the other choice
zero and one.

212
00:12:42 --> 00:12:43
You see my system.

213
00:12:43 --> 00:12:45
So let me repeat the system.

214
00:12:45 --> 00:12:49
This is the algorithm that you,
you just learned to do.

215
00:12:49 --> 00:12:50
Do elimination.

216
00:12:50 --> 00:12:54
Decide which are the pivot
columns and which are the free

217
00:12:54 --> 00:12:55
columns.

218
00:12:55 --> 00:12:58.72
That tells you that,
that variables one and three

219
00:12:58.72 --> 00:13:01.72
are pivot variables,
two and four are free

220
00:13:01.72 --> 00:13:03
variables.

221
00:13:03 --> 00:13:09
Then those free variables,
you assign them -- you give one

222
00:13:09 --> 00:13:16
of them the value one and the
others the value zero -- in this

223
00:13:16 --> 00:13:22
case, we had a one and a zero --
and complete the solution.

224
00:13:22 --> 00:13:26
And you do --
you give the other one the

225
00:13:26 --> 00:13:27
value one and zero.

226
00:13:27 --> 00:13:29
And now complete the solution.

227
00:13:29 --> 00:13:31
So let's complete that
solution.

228
00:13:31 --> 00:13:35
I'm looking for a vector in the
null space, and it's absolutely

229
00:13:35 --> 00:13:38
going to be different from this
guy, because,

230
00:13:38 --> 00:13:42
because any multiple of that
zero is never going to give the

231
00:13:42 --> 00:13:43
one.

232
00:13:43 --> 00:13:48
So I have somebody new in the
null space, and let me finish it

233
00:13:48 --> 00:13:48
out.

234
00:13:48 --> 00:13:50
What's x3 here?

235
00:13:50 --> 00:13:53
So we're going by back
substitution,

236
00:13:53 --> 00:13:55
looking at this equation.

237
00:13:55 --> 00:13:59
Now x4 we've changed,
we're doing the other

238
00:13:59 --> 00:14:03
possibility, where x2 is zero
and x4 is one.

239
00:14:03 --> 00:14:07
So x3 will happen to be minus
two.

240
00:14:07 --> 00:14:12
And now what do I get for that
first equation?

241
00:14:12 --> 00:14:13
x1 -- let's see.

242
00:14:13 --> 00:14:19
Two x3s is minus four plus two
-- do I get a two there?

243
00:14:19 --> 00:14:20
Perhaps, yeah.

244
00:14:20 --> 00:14:23
Two for x1, minus four,
and two.

245
00:14:23 --> 00:14:24
OK.

246
00:14:24 --> 00:14:26
That's in the null space.

247
00:14:26 --> 00:14:31
What does that say about the
columns?

248
00:14:31 --> 00:14:36
That says that two of this
column minus two of this column

249
00:14:36 --> 00:14:39
plus this column gives zero,
which it does.

250
00:14:39 --> 00:14:44.44
Two of that minus two of that
and one of that gives the zero

251
00:14:44.44 --> 00:14:45
column.

252
00:14:45 --> 00:14:49
OK, now, now I've found another
vector in the null space.

253
00:14:49 --> 00:14:53
Now we're ready to tell me the
whole null space.

254
00:14:53 --> 00:14:57
What are all the solutions to
Ax=0?

255
00:14:57 --> 00:15:02
I've got this guy and when I
have him, what else is,

256
00:15:02 --> 00:15:07
goes into the null space along
with that?

257
00:15:07 --> 00:15:11
These are my two special
solutions.

258
00:15:11 --> 00:15:16
I call them special -- I just
invented that name.

259
00:15:16 --> 00:15:19
Special solutions.

260
00:15:19 --> 00:15:24
What's special about them is
the special numbers I gave to

261
00:15:24 --> 00:15:28
the free variables,
the values zero and one for the

262
00:15:28 --> 00:15:29.55
free variables.

263
00:15:29.55 --> 00:15:34
I could have given the free
variables any values and got

264
00:15:34 --> 00:15:36
vectors in the null space.

265
00:15:36 --> 00:15:41
But this was a good way to be
sure I got t- got everybody.

266
00:15:41 --> 00:15:45
OK, so once I have him,
I also have any multiple,

267
00:15:45 --> 00:15:46
right?

268
00:15:46 --> 00:15:51
So I could take any multiple of
that and it's in the null space.

269
00:15:51 --> 00:15:55
And now what else -- I left a
little space for what?

270
00:15:55 --> 00:15:56
What -- a plus sign.

271
00:15:56 --> 00:15:59
I can take any combination.

272
00:15:59 --> 00:16:02
Here is a line of vectors in
the null space.

273
00:16:02 --> 00:16:05
A bunch of solutions.

274
00:16:05 --> 00:16:11.76
Would you rather I say in the
null space or would you rather I

275
00:16:11.76 --> 00:16:14.53
say, OK, I'm solving Ax=0?

276
00:16:14.53 --> 00:16:17
Well, really I'm solving Ux=0.

277
00:16:17 --> 00:16:22
Well, OK, let me put in that
crucial plus sign.

278
00:16:22 --> 00:16:28
I'm taking all the combinations
of my two special solutions.

279
00:16:28 --> 00:16:32
That's my conclusion there.

280
00:16:32 --> 00:16:36
The null space contains,
contains exactly all the

281
00:16:36 --> 00:16:39
combinations of the special
solutions.

282
00:16:39 --> 00:16:42
And how many special solutions
are there?

283
00:16:42 --> 00:16:44
There's one for every free
variable.

284
00:16:44 --> 00:16:47.54
And how many free variables are
there?

285
00:16:47.54 --> 00:16:52
Oh, I mean, we can see all the
whole picture now.

286
00:16:52 --> 00:16:55
If the rank R was two,
this is the,

287
00:16:55 --> 00:17:00
this is the number of pivot
variables, right,

288
00:17:00 --> 00:17:03
because it counted the pivots.

289
00:17:03 --> 00:17:06
So how many free variables?

290
00:17:06 --> 00:17:10
Well, you know it's two,
right?

291
00:17:10 --> 00:17:14
What is it in -- for a matrix
that's m rows,

292
00:17:14 --> 00:17:21
n columns, n variables that
means, with rank r?

293
00:17:21 --> 00:17:27
How many free variables have we
got left?

294
00:17:27 --> 00:17:36
If r of the variables are pivot
variables, we have n-r -- in

295
00:17:36 --> 00:17:42
this case four minus two -- free
variables.

296
00:17:42 --> 00:17:52
Do you see that first of all we
get clean answers here?

297
00:17:52 --> 00:17:56
We get r pivot variables -- so
there really were r equations

298
00:17:56 --> 00:17:57
here.

299
00:17:57 --> 00:18:01.09
There looked like three
equations, but there were really

300
00:18:01.09 --> 00:18:03
only two independent equations.

301
00:18:03 --> 00:18:07
And there were n-r variables
that we could choose freely,

302
00:18:07 --> 00:18:10
and we gave them those special
zero one values,

303
00:18:10 --> 00:18:13
and we got the special
solutions.

304
00:18:13 --> 00:18:14
OK.

305
00:18:14 --> 00:18:20
For me -- we could stop at that
point.

306
00:18:20 --> 00:18:28
That gives you a complete
algorithm for finding all the

307
00:18:28 --> 00:18:33
solutions to A x equals zero.

308
00:18:33 --> 00:18:33
OK.

309
00:18:33 --> 00:18:41
Again, you do elimination --
going onward when a column,

310
00:18:41 --> 00:18:46
when there's nothing to be done
on one column,

311
00:18:46 --> 00:18:48.03
you just continue.

312
00:18:48.03 --> 00:18:52
There's this number r,
the number of pivots,

313
00:18:52 --> 00:18:56
is crucial, and it leaves n-r
free variables,

314
00:18:56 --> 00:19:00
which you give values zero and
one to.

315
00:19:00 --> 00:19:04
I would like to take one more
step.

316
00:19:04 --> 00:19:09
I would like to clean up this
matrix even more.

317
00:19:09 --> 00:19:14
So now I'm going to go to --
this is in its,

318
00:19:14 --> 00:19:20
this is in echelon form,
upper triangular if you like.

319
00:19:20 --> 00:19:26
I want to go one more step to
make it as good as it can be.

320
00:19:26 --> 00:19:32
OK, so now I'm going to speak
about the reduced row echelon

321
00:19:32 --> 00:19:34.3
form.

322
00:19:34.3 --> 00:19:41
OK, so now I'm going to speak
about the matrix R,

323
00:19:41 --> 00:19:46
which is the reduced row
echelon form.

324
00:19:46 --> 00:19:50
So what does that mean?

325
00:19:50 --> 00:19:57.83
That means I just -- I can,
I can work harder on U.

326
00:19:57.83 --> 00:20:04
So let me start,
let me suppose I got as far as

327
00:20:04 --> 00:20:08
U, which was good.

328
00:20:08 --> 00:20:12
Notice how that row of zeros
appeared.

329
00:20:12 --> 00:20:16
I didn't comment on that,
but I should have.

330
00:20:16 --> 00:20:21.08
That row of zeros up here is
because row three was a

331
00:20:21.08 --> 00:20:26
combination of rows one and two,
and elimination discovered that

332
00:20:26 --> 00:20:27
fact.

333
00:20:27 --> 00:20:33
When we get a row of zeros,
that's telling us that the --

334
00:20:33 --> 00:20:40
original row that was there was
a combination of other rows,

335
00:20:40 --> 00:20:44
and elimination knocked it out.

336
00:20:44 --> 00:20:46
OK, so we got this far.

337
00:20:46 --> 00:20:51.1
Now how can I clean that up
further?

338
00:20:51.1 --> 00:20:54.63
I can do, elimination upwards.

339
00:20:54.63 --> 00:20:58
I can get zero above the
pivots.

340
00:20:58 --> 00:21:04
So this reduced row echelon
form has zeros above and below

341
00:21:04 --> 00:21:06.28
the pivots.

342
00:21:06.28 --> 00:21:09
So let me do that.

343
00:21:09 --> 00:21:14
So now I'll subtract one of
this from the row above.

344
00:21:14 --> 00:21:19
That will leave a zero and a
minus two in there.

345
00:21:19 --> 00:21:21
And that's good.

346
00:21:21 --> 00:21:25
OK, and I can clean it up even
one more step.

347
00:21:25 --> 00:21:31
I can make the pivots -- the
pivots I'm going to make equal

348
00:21:31 --> 00:21:37
to one, because I can divide
equation two by the pivot.

349
00:21:37 --> 00:21:41
That won't change the
solutions.

350
00:21:41 --> 00:21:43
So let me do that.

351
00:21:43 --> 00:21:47
And then I really -- I'm ready
to stop.

352
00:21:47 --> 00:21:51
One, two, zero,
minus two, zero,

353
00:21:51 --> 00:21:52
zero, one, two.

354
00:21:52 --> 00:21:59
I divided the second equation
by two, because now I have a one

355
00:21:59 --> 00:22:04
in the pivot and zeros below.

356
00:22:04 --> 00:22:04
OK.

357
00:22:04 --> 00:22:07
This is my matrix R.

358
00:22:07 --> 00:22:13
I guess I'm hoping that you
could now execute the whole

359
00:22:13 --> 00:22:15
algorithm.

360
00:22:15 --> 00:22:22
Matlab will do it immediately
with the command -- reduced row

361
00:22:22 --> 00:22:24
echelon form of A.

362
00:22:24 --> 00:22:31
So if I input that original
matrix A and then I write,

363
00:22:31 --> 00:22:35
then I type that command,
press return,

364
00:22:35 --> 00:22:40
that matrix will appear.

365
00:22:40 --> 00:22:47
That's the reduced row echelon
form, and it's got all the

366
00:22:47 --> 00:22:51
information as clear as can be.

367
00:22:51 --> 00:22:55
What, what information has it
got?

368
00:22:55 --> 00:23:02
Well, of course it immediately
tells me the pivot rows,

369
00:23:02 --> 00:23:07
pivot rows, one and two,
pivot columns,

370
00:23:07 --> 00:23:10
one and three.

371
00:23:10 --> 00:23:15.21
And in fact it's got the
identity matrix in there,

372
00:23:15.21 --> 00:23:15
right?

373
00:23:15 --> 00:23:20
It's, it's got zeros above and
below the pivots,

374
00:23:20 --> 00:23:25
and the pivots are one,
so it's, so it's got a -- so

375
00:23:25 --> 00:23:31
notice the two by two identity
matrix that's sitting in the

376
00:23:31 --> 00:23:36
pivot rows and pivot columns.
it's I in the pivot rows and

377
00:23:36 --> 00:23:38
columns.

378
00:23:38 --> 00:23:42
It's got zero rows below.

379
00:23:42 --> 00:23:49
Those are always indicating
that original rows were,

380
00:23:49 --> 00:23:53
were combinations of other
rows.

381
00:23:53 --> 00:23:58
So we really only had two rows
there.

382
00:23:58 --> 00:24:04
And now it also -- so there's
the identity.

383
00:24:04 --> 00:24:10
Now it's also got its free
columns.

384
00:24:10 --> 00:24:15
And, they're cleaned up as much
as possible.

385
00:24:15 --> 00:24:20
Actually, actually it's now so
cleaned up that the special

386
00:24:20 --> 00:24:26
solutions, I can read off -- you
remember I went through the

387
00:24:26 --> 00:24:31
steps of computing this -- doing
back substitution?

388
00:24:31 --> 00:24:35
Let me, let me,
instead of that system,

389
00:24:35 --> 00:24:38.09
let me take this improved
system.

390
00:24:38.09 --> 00:24:41.71
So I'm going to use these
numbers, right.

391
00:24:41.71 --> 00:24:44
In these equations,
what did I do?

392
00:24:44 --> 00:24:49
I divided this equation by two
and, oh yeah and I had

393
00:24:49 --> 00:24:54.2
subtracted two of this,
which knocked out this guy and

394
00:24:54.2 --> 00:24:57
made that a minus sign.

395
00:24:57 --> 00:25:03
Is that what -- I've now
written Rx equals zero.

396
00:25:03 --> 00:25:08
Now I guess I'm hoping
everybody in this room

397
00:25:08 --> 00:25:15
understands the solutions to the
original A x equals zero,

398
00:25:15 --> 00:25:19
the midway, halfway,
U x equals zero,

399
00:25:19 --> 00:25:26
and the final R x equals zero
are all the same.

400
00:25:26 --> 00:25:32
Because going from one of those
to another one I didn't mess up.

401
00:25:32 --> 00:25:38
I just multiplied equations and
subtracted from other equations,

402
00:25:38 --> 00:25:40
which I'm allowed to do.

403
00:25:40 --> 00:25:40
OK.

404
00:25:40 --> 00:25:46
But my point is that now if I
do this free variables and back

405
00:25:46 --> 00:25:49
substitution,
it's just, the numbers are

406
00:25:49 --> 00:25:51
there.

407
00:25:51 --> 00:25:57
When I let x -- so in this guy,
I let x2 be one and x4 be zero.

408
00:25:57 --> 00:26:00
I, I guess, what I seeing here?

409
00:26:00 --> 00:26:05
Let me, let me sort of separate
this out here.

410
00:26:05 --> 00:26:09
I'm seeing in the pivot,
in the pivot columns,

411
00:26:09 --> 00:26:13
if I, if I put the pivot
columns here,

412
00:26:13 --> 00:26:15
I'm seeing those.

413
00:26:15 --> 00:26:21
And I'm -- in the free columns
I'm seeing -- what I seeing in

414
00:26:21 --> 00:26:23
the free columns?

415
00:26:23 --> 00:26:28
A two, zero in that first free
column, the x2 column,

416
00:26:28 --> 00:26:31
and a minus two,
two in the fourth column,

417
00:26:31 --> 00:26:33
the other free column.

418
00:26:33 --> 00:26:39
And the row of zeros below,
which of course have no -- that

419
00:26:39 --> 00:26:43
equation is zero equals zero.

420
00:26:43 --> 00:26:44
That's satisfied.

421
00:26:44 --> 00:26:46
Here's my point.

422
00:26:46 --> 00:26:48
That when I do back
substitution,

423
00:26:48 --> 00:26:52
these numbers are exactly what
shows up -- oh,

424
00:26:52 --> 00:26:55
their signs get switched.

425
00:26:55 --> 00:26:58
I was going to say those
numbers, two,

426
00:26:58 --> 00:27:02
minus two, zero,
two, can I just circle the --

427
00:27:02 --> 00:27:05
this is the free part of the
matrix.

428
00:27:05 --> 00:27:08
This is the identity part.

429
00:27:08 --> 00:27:12
This is the free part,
maybe I'll call it F.

430
00:27:12 --> 00:27:15
This, of course,
I call I, because it's the

431
00:27:15 --> 00:27:16
identity.

432
00:27:16 --> 00:27:19
The free part is a,
I mean, I'm just doing back

433
00:27:19 --> 00:27:20
substitution.

434
00:27:20 --> 00:27:24
And those free numbers will
show up, with a minus sign,

435
00:27:24 --> 00:27:29
because they pop onto the other
side of the equation --

436
00:27:29 --> 00:27:34
so I see minus two,
zero, and I see two,

437
00:27:34 --> 00:27:35
minus two.

438
00:27:35 --> 00:27:38
So that wasn't magic.

439
00:27:38 --> 00:27:40
It had to happen.

440
00:27:40 --> 00:27:44.87
Let me, show you clearly why it
happened.

441
00:27:44.87 --> 00:27:52
OK, so that's -- this is what
I'm interested in here.

442
00:27:52 --> 00:27:55
And now let me,
let me just,

443
00:27:55 --> 00:27:59
like, do it,
do it for -- let's suppose

444
00:27:59 --> 00:28:06.22
we've, we've got to -- let's
suppose we've got this system

445
00:28:06.22 --> 00:28:09
already in, in rref form.

446
00:28:09 --> 00:28:14
So my matrix R is -- what does
it look like?

447
00:28:14 --> 00:28:20
OK, and I'll -- let me pretend
that the pivot columns come

448
00:28:20 --> 00:28:27
first and then whatever's in the
free columns.

449
00:28:27 --> 00:28:34
And there might be some zero
rows below.

450
00:28:34 --> 00:28:45
There's a typical -- a pretty
typical reduced row echelon

451
00:28:45 --> 00:28:46
form.

452
00:28:46 --> 00:28:50
You see what's typical.

453
00:28:50 --> 00:28:55
It's got -- this is r by r.

454
00:28:55 --> 00:29:01
This is r pivot rows.

455
00:29:01 --> 00:29:04
This is r pivot columns.

456
00:29:04 --> 00:29:08
And here are n-r free columns.

457
00:29:08 --> 00:29:09
OK.

458
00:29:09 --> 00:29:14
Tell me what are the special
solutions?

459
00:29:14 --> 00:29:18
What are the -- what's x?

460
00:29:18 --> 00:29:25
If I want to solve R x equals
zero -- in fact,

461
00:29:25 --> 00:29:33
let me -- I'm really going to,
do the whole --

462
00:29:33 --> 00:29:37
since these -- this is now
block matrices,

463
00:29:37 --> 00:29:42
I might as well do all of the
special solutions at once.

464
00:29:42 --> 00:29:48
So I want to solve R x equals
zero, and I'll have some special

465
00:29:48 --> 00:29:49
solutions.

466
00:29:49 --> 00:29:54
Let me, actually -- can I do
them all at once?

467
00:29:54 --> 00:29:58
I'm going to create a null
space matrix,

468
00:29:58 --> 00:29:59
OK.

469
00:29:59 --> 00:30:00
A matrix.

470
00:30:00 --> 00:30:07
Its, its, its columns are the
special -- the columns are the

471
00:30:07 --> 00:30:09
special solutions.

472
00:30:09 --> 00:30:15
This is, I'm making it sound
harder, it's going to be totally

473
00:30:15 --> 00:30:16
easy.

474
00:30:16 --> 00:30:19
N will be this null space
matrix.

475
00:30:19 --> 00:30:23
I want R N to be the zero
matrix.

476
00:30:23 --> 00:30:30
These columns of N are supposed
to multipl- to get multiplied by

477
00:30:30 --> 00:30:34
R and give zero columns.

478
00:30:34 --> 00:30:36
So what N will do the job?

479
00:30:36 --> 00:30:42
Let me put -- I'm going to put
the identity in the free

480
00:30:42 --> 00:30:47
variable part and then minus F
will show up in the pivot

481
00:30:47 --> 00:30:52
variables, just the way it did
in that example.

482
00:30:52 --> 00:30:55
There we had the identity and
F.

483
00:30:55 --> 00:30:59
Here -- in the special
solution.

484
00:30:59 --> 00:31:05
So these columns are -- there's
the matrix of special solutions.

485
00:31:05 --> 00:31:09
And actually,
there -- so there's a Matlab

486
00:31:09 --> 00:31:13
command or a teaching code
command, NULL -- N equal,

487
00:31:13 --> 00:31:18
so this is the -- produces the
null basis, the null space

488
00:31:18 --> 00:31:22
matrix, NULL of A,
and there it is.

489
00:31:22 --> 00:31:27
And how does that command
actually work?

490
00:31:27 --> 00:31:33.88
It uses Matlab to compute R,
then it picks out the pivot

491
00:31:33.88 --> 00:31:40
variables, the free variables,
puts, puts ones and zeros in

492
00:31:40 --> 00:31:46.15
for the free variables,
and copies out the pivot

493
00:31:46.15 --> 00:31:48
variables.

494
00:31:48 --> 00:31:55
It, it does back substitution,
but back substitution for this

495
00:31:55 --> 00:31:58
system is totally simple.

496
00:31:58 --> 00:32:01
What is this system?

497
00:32:01 --> 00:32:03
R x equals zero.

498
00:32:03 --> 00:32:09
So this is R is I F,
and x is the pivot variables

499
00:32:09 --> 00:32:16
and the free variables,
and it's supposed to give zero.

500
00:32:16 --> 00:32:19
So what does that mean?

501
00:32:19 --> 00:32:24
That means that the pivot
variables plus F times the free

502
00:32:24 --> 00:32:26
variables give zero.

503
00:32:26 --> 00:32:32.56
So let me put F times the free
variables on the other side.

504
00:32:32.56 --> 00:32:36
I get minus F times the free
variables.

505
00:32:36 --> 00:32:41
There's my, equation,
as simple as it can be.

506
00:32:41 --> 00:32:47
That's what back substitution
comes to when I've reduced and

507
00:32:47 --> 00:32:52
reduced and reduced this system
to the, to the best form,

508
00:32:52 --> 00:32:52
OK.

509
00:32:52 --> 00:32:57
And, then if the free
variables, I make this special

510
00:32:57 --> 00:33:02
choice of the identity,
then the pivot variables are

511
00:33:02 --> 00:33:02
minus F.

512
00:33:02 --> 00:33:05
OK, can I do,
another example?

513
00:33:05 --> 00:33:09
Could you do another example?

514
00:33:09 --> 00:33:13
Can I -- let me just take
another matrix and,

515
00:33:13 --> 00:33:18
and let's go through this
algorithm once more,

516
00:33:18 --> 00:33:18
OK.

517
00:33:18 --> 00:33:19
Here we go.

518
00:33:19 --> 00:33:24
Here's a blackboard for another
matrix, OK.

519
00:33:24 --> 00:33:29
So I'll call the matrix A
again, but let me make it --

520
00:33:29 --> 00:33:34
yeah, how big shall we make it
this time?

521
00:33:34 --> 00:33:37
Why don't I do this?

522
00:33:37 --> 00:33:40
Just for the heck of it.

523
00:33:40 --> 00:33:49
Let me take the transpose of
this A and see what happens to

524
00:33:49 --> 00:33:49
that.

525
00:33:49 --> 00:33:55
Two four six eight and three
six eight ten.

526
00:33:55 --> 00:34:03
Before we do the calculations,
tell me what's coming?

527
00:34:03 --> 00:34:10
How many pivot variables do you
expect here?

528
00:34:10 --> 00:34:14
How many columns are going to
have pivots?

529
00:34:14 --> 00:34:18
How many -- we have three
columns in that matrix,

530
00:34:18 --> 00:34:22
but are we going to,
are we going to have three

531
00:34:22 --> 00:34:23
pivots?

532
00:34:23 --> 00:34:28
No, because this third columns
is the sum of the first two

533
00:34:28 --> 00:34:28
columns.

534
00:34:28 --> 00:34:33
I'm totally expecting,
totally expecting that these

535
00:34:33 --> 00:34:38
will be pivot columns --
because they're independent,

536
00:34:38 --> 00:34:42.57
but that this third guy,
the third column,

537
00:34:42.57 --> 00:34:47
which is dependent on the first
two, is going to be a free

538
00:34:47 --> 00:34:48
column.

539
00:34:48 --> 00:34:51
Elimination better discover
that.

540
00:34:51 --> 00:34:55
And elimination will also
straighten out the rows,

541
00:34:55 --> 00:35:00
dependent rows and some
independent rows.

542
00:35:00 --> 00:35:05
What's the, what's the row
reduced echelon form for this?

543
00:35:05 --> 00:35:07
Let's just do it,
OK.

544
00:35:07 --> 00:35:10
So, so that's the first pivot.

545
00:35:10 --> 00:35:15
Two times that away from that
gives me a row of zeros.

546
00:35:15 --> 00:35:21
Two times that away from that
gives me a zero two two.

547
00:35:21 --> 00:35:27
And two times that away from
that gives me a zero four four.

548
00:35:27 --> 00:35:30
OK, first column is straight.

549
00:35:30 --> 00:35:34
First variable is a pivot
variable.

550
00:35:34 --> 00:35:35
No problem.

551
00:35:35 --> 00:35:37
On to the second column.

552
00:35:37 --> 00:35:41
I look at the second pivot,
it's a zero.

553
00:35:41 --> 00:35:44
I look below it.

554
00:35:44 --> 00:35:45
There's a two.

555
00:35:45 --> 00:35:48
OK, I do a row exchange.

556
00:35:48 --> 00:35:50
So this zero is now there.

557
00:35:50 --> 00:35:55
I now have a perfectly good
pivot, and I use it.

558
00:35:55 --> 00:36:00
OK, and I subtract two of that
row away from this row.

559
00:36:00 --> 00:36:04
All right if I do it like that?

560
00:36:04 --> 00:36:06
I've got to the form U now.

561
00:36:06 --> 00:36:09
This was my A.

562
00:36:09 --> 00:36:10
Now there's my U.

563
00:36:10 --> 00:36:13
I can see now -- I have to
stop, right?

564
00:36:13 --> 00:36:16
I would go on to the third
column.

565
00:36:16 --> 00:36:18
I should have tried.

566
00:36:18 --> 00:36:20
I quit, but without trying.

567
00:36:20 --> 00:36:22
I shouldn't have done that.

568
00:36:22 --> 00:36:26.79
On to the third column,
look at the pivot position.

569
00:36:26.79 --> 00:36:29
It's got a zero in it.

570
00:36:29 --> 00:36:31
Look below, all zeros.

571
00:36:31 --> 00:36:34
Now I'm entitled to stop,
OK.

572
00:36:34 --> 00:36:36
So the rank is two again.

573
00:36:36 --> 00:36:39
What about the null space?

574
00:36:39 --> 00:36:44.16
How many special solutions are
there this time?

575
00:36:44.16 --> 00:36:48
How many special solutions for
this matrix?

576
00:36:48 --> 00:36:54
I've got -- and which are the
free variables and which are the

577
00:36:54 --> 00:36:58
pivot variables and so on?

578
00:36:58 --> 00:37:02
Pivot columns,
I've got two pivot columns,

579
00:37:02 --> 00:37:04
and that's no accident.

580
00:37:04 --> 00:37:10
The number of pivot columns for
a matrix A, that's a great fact,

581
00:37:10 --> 00:37:16
that the number of pivot
columns for A and A transpose

582
00:37:16 --> 00:37:18
are the same.

583
00:37:18 --> 00:37:22
And then I have a free column.

584
00:37:22 --> 00:37:24
There's a free column.

585
00:37:24 --> 00:37:30
One free column,
because the count is three

586
00:37:30 --> 00:37:31
minus two.

587
00:37:31 --> 00:37:36
Three minus two gives me one
free column.

588
00:37:36 --> 00:37:42
OK, so now let me solve,
what's in the null space.

589
00:37:42 --> 00:37:46
OK, so how do I -- let's see.

590
00:37:46 --> 00:37:52
These vectors have length
three.

591
00:37:52 --> 00:37:55
They only have three
components.

592
00:37:55 --> 00:38:00
I'm making too much space for
the, to write x.

593
00:38:00 --> 00:38:05
x has just got three
components, and what are they?

594
00:38:05 --> 00:38:08
I'm looking for the null space.

595
00:38:08 --> 00:38:10
OK, so how do I start?

596
00:38:10 --> 00:38:16
I give the free variable some
convenient value.

597
00:38:16 --> 00:38:18.25
And what's that?

598
00:38:18.25 --> 00:38:19.82
I set it to one.

599
00:38:19.82 --> 00:38:22
I set the free variable to one.

600
00:38:22 --> 00:38:28
If I set the free variable to
zero and solve for the pivot

601
00:38:28 --> 00:38:32
variables, I'll get all zeros:
no progress.

602
00:38:32 --> 00:38:37
But by setting the free
variable to one -- you see w- my

603
00:38:37 --> 00:38:44
two equations now are --
my equations are x1+ 2x2+ 3 x3

604
00:38:44 --> 00:38:48
is zero, that's my first
equation.

605
00:38:48 --> 00:38:54
And my second equation is now
2x2+2x3 equals zero.

606
00:38:54 --> 00:38:55
And, OK.

607
00:38:55 --> 00:39:00.31
So if x3 is one,
then x2 is minus one.

608
00:39:00.31 --> 00:39:06
And if x3 is one and x2 is
minus one, then maybe x1 is

609
00:39:06 --> 00:39:07
minus one.

610
00:39:07 --> 00:39:13.74
And actually I go back to check
now.

611
00:39:13.74 --> 00:39:17
I don't, like -- I did a quick
calculation mentally.

612
00:39:17 --> 00:39:20
Can I mentally do a quick
check?

613
00:39:20 --> 00:39:23
That matrix,
that solution x says that minus

614
00:39:23 --> 00:39:27
this column minus this column
plus this one is the zero

615
00:39:27 --> 00:39:28
column.

616
00:39:28 --> 00:39:29
And it is.

617
00:39:29 --> 00:39:33
Minus that minus that plus that
is zero.

618
00:39:33 --> 00:39:35
So that's in the null space.

619
00:39:35 --> 00:39:39
And now you can tell me what
else is in the null space.

620
00:39:39 --> 00:39:42
What's, what's the whole null
space now?

621
00:39:42 --> 00:39:44
I multiply by C,
right.

622
00:39:44 --> 00:39:46
The whole null space is a line.

623
00:39:46 --> 00:39:48
So that's the description.

624
00:39:48 --> 00:39:53
You know, if I ask you on a
homework or a quiz or the final

625
00:39:53 --> 00:39:56
what --
give me, describe,

626
00:39:56 --> 00:40:00
tell me the null space,
find the null space of this

627
00:40:00 --> 00:40:02
matrix, you can take those
steps.

628
00:40:02 --> 00:40:05
And that's the answer I'm
looking for.

629
00:40:05 --> 00:40:10
And I'm looking for that C too,
because that's telling me that

630
00:40:10 --> 00:40:15.42
you're remembering that it's a
whole space and not just one

631
00:40:15.42 --> 00:40:16
vector.

632
00:40:16 --> 00:40:21
Later I will ask you for a
basis for the null space.

633
00:40:21 --> 00:40:23
Then I just want this vector.

634
00:40:23 --> 00:40:28
But if I ask for the whole null
space, then there's the whole

635
00:40:28 --> 00:40:30
line through that vector.

636
00:40:30 --> 00:40:34
OK, now one more natural thing
to do with this example,

637
00:40:34 --> 00:40:38
right, is keep going to the
reduced matrix,

638
00:40:38 --> 00:40:39
R.

639
00:40:39 --> 00:40:42
So can I push onwards to R?

640
00:40:42 --> 00:40:46
That should be quick,
but let's just practice.

641
00:40:46 --> 00:40:48.71
Let me keep going to R.

642
00:40:48.71 --> 00:40:51
OK, so what do I do here?

643
00:40:51 --> 00:40:57.01
I subtract -- I clear out above
the pivot, so I subtract that

644
00:40:57.01 --> 00:41:02
from that, that's one zero one
is left, right?

645
00:41:02 --> 00:41:07
When I subtracted this row from
this it produced a zero above

646
00:41:07 --> 00:41:08
this pivot.

647
00:41:08 --> 00:41:12
And now I want that pivot to be
a one.

648
00:41:12 --> 00:41:16
So for the R matrix,
I'll divide this equation by

649
00:41:16 --> 00:41:21
two, and of course these zero,
zeros are great,

650
00:41:21 --> 00:41:22
they don't change.

651
00:41:22 --> 00:41:23
There's R.

652
00:41:23 --> 00:41:25
That's R.

653
00:41:25 --> 00:41:27
You see what R is?

654
00:41:27 --> 00:41:32
You see the identity matrix
sitting up here?

655
00:41:32 --> 00:41:37
You see the free part F,
the F part here?

656
00:41:37 --> 00:41:40
And you see the zeros below.

657
00:41:40 --> 00:41:43
This is I F zero zero.

658
00:41:43 --> 00:41:45
And what's the x?

659
00:41:45 --> 00:41:52
The x has the identity --
well, it's only a single number

660
00:41:52 --> 00:41:55
one, but it's the identity
matrix in the free,

661
00:41:55 --> 00:41:57
in the free part.

662
00:41:57 --> 00:42:01
And what does it have in the
pivot variables?

663
00:42:01 --> 00:42:03.72
What did back substitution
give?

664
00:42:03.72 --> 00:42:05
It gave minus these guys.

665
00:42:05 --> 00:42:10
You see that what this is is
any multiple of -- this is the

666
00:42:10 --> 00:42:15
identity there,
and this is minus F here.

667
00:42:15 --> 00:42:18
This is our null space matrix N
for this.

668
00:42:18 --> 00:42:24
Our, our null space matrix is
the guy whose columns are the

669
00:42:24 --> 00:42:25
special solutions.

670
00:42:25 --> 00:42:31
So their free variables have
the special values one and,

671
00:42:31 --> 00:42:33
pivot variables have minus F.

672
00:42:33 --> 00:42:37
So do you see,
though, how the minus F just

673
00:42:37 --> 00:42:43
automatically shows up in the
special solutions.

674
00:42:43 --> 00:42:45
That's it really.

675
00:42:45 --> 00:42:51
I don't think there's anything
more I can say about A x equals

676
00:42:51 --> 00:42:52
zero.

677
00:42:52 --> 00:42:58
There's more I can say about A
x equal b, but that'll be on

678
00:42:58 --> 00:42:59
Friday.

679
00:42:59 --> 00:43:04
OK, so that's,
that's the null space.

680
00:43:04 --> 00:43:07
Thanks.