• Title: Lebesgue Integral

  • Series: Measure Theory

  • YouTube-Title: Measure Theory 6 | Lebesgue Integral

  • Bright video: https://youtu.be/TG67nsccqeQ

  • Dark video: https://youtu.be/f4FqnwEgR98

  • Quiz: Test your knowledge

  • PDF: Download PDF version of the bright video

  • Dark-PDF: Download PDF version of the dark video

  • Print-PDF: Download printable PDF version

  • Thumbnail (bright): Download PNG

  • Thumbnail (dark): Download PNG

  • Subtitle on GitHub: mt06_sub_eng.srt

  • Other languages: German version

  • Timestamps (n/a)
  • Subtitle in English

    1 00:00:00,579 –> 00:00:02,170 Hello and welcome

    2 00:00:02,180 –> 00:00:02,720 back.

    3 00:00:02,730 –> 00:00:04,519 First, let me thank all the

    4 00:00:04,530 –> 00:00:06,219 nice people that support

    5 00:00:06,230 –> 00:00:07,760 this channel on Steady.

    6 00:00:08,850 –> 00:00:10,539 And now let’s continue with

    7 00:00:10,550 –> 00:00:12,439 measure theory namely

    8 00:00:12,449 –> 00:00:14,069 with part six.

    9 00:00:15,989 –> 00:00:17,420 And we will finally talk

    10 00:00:17,430 –> 00:00:19,190 about the Lebesgue integral but

    11 00:00:19,200 –> 00:00:20,639 first about the Lebesgue

    12 00:00:20,649 –> 00:00:22,399 integral of so

    13 00:00:22,409 –> 00:00:24,280 called-step functions.

    14 00:00:25,090 –> 00:00:26,920 So we will learn how to

    15 00:00:26,930 –> 00:00:28,799 integrate functions that

    16 00:00:28,809 –> 00:00:30,120 are defined on an

    17 00:00:30,129 –> 00:00:31,840 abstract measure space.

    18 00:00:33,020 –> 00:00:34,619 As a short recap, a

    19 00:00:34,630 –> 00:00:36,240 measure space is

    20 00:00:36,250 –> 00:00:37,709 nothing more than a

    21 00:00:37,720 –> 00:00:38,400 triple.

    22 00:00:39,110 –> 00:00:39,330 There.

    23 00:00:39,340 –> 00:00:40,779 We have a set X,

    24 00:00:41,209 –> 00:00:42,770 a sigma algebra A

    25 00:00:42,930 –> 00:00:44,470 and also a measure

    26 00:00:44,479 –> 00:00:45,009 mu.

    27 00:00:45,830 –> 00:00:47,779 This means that X could

    28 00:00:47,790 –> 00:00:49,080 be any set

    29 00:00:50,049 –> 00:00:51,979 but A is a special

    30 00:00:51,990 –> 00:00:53,580 collection of subsets of

    31 00:00:53,590 –> 00:00:54,049 X.

    32 00:00:54,529 –> 00:00:56,189 And the third ingredient,

    33 00:00:56,360 –> 00:00:58,040 the measure mu itself is

    34 00:00:58,049 –> 00:00:59,750 indeed a map

    35 00:01:00,259 –> 00:01:02,139 where the domain is the sigma

    36 00:01:02,150 –> 00:01:03,880 algebra A and the co-

    37 00:01:03,889 –> 00:01:05,290 domain is the

    38 00:01:05,300 –> 00:01:06,980 interval included

    39 00:01:06,989 –> 00:01:08,580 zero and also

    40 00:01:08,589 –> 00:01:10,220 included the symbol

    41 00:01:10,230 –> 00:01:10,910 infinity.

    42 00:01:11,860 –> 00:01:13,580 Now, with respect to

    43 00:01:13,589 –> 00:01:15,519 this abstract measure

    44 00:01:15,529 –> 00:01:16,870 space, we want to

    45 00:01:16,879 –> 00:01:18,720 integrate some special

    46 00:01:18,730 –> 00:01:19,510 functions.

    47 00:01:20,470 –> 00:01:22,139 Indeed what we need are

    48 00:01:22,150 –> 00:01:24,089 measurable maps defined in

    49 00:01:24,099 –> 00:01:25,970 the last video I

    50 00:01:25,980 –> 00:01:27,559 will use the letter F for

    51 00:01:27,569 –> 00:01:29,529 such maps that start

    52 00:01:29,540 –> 00:01:31,529 with X and go into

    53 00:01:31,540 –> 00:01:32,730 the real number line.

    54 00:01:33,709 –> 00:01:35,610 Now you should not forget

    55 00:01:35,620 –> 00:01:37,519 that we have a Sigma algebra

    56 00:01:37,529 –> 00:01:39,089 on the left namely

    57 00:01:39,099 –> 00:01:40,970 A and also a Sigma

    58 00:01:40,980 –> 00:01:42,150 algebra on the right.

    59 00:01:42,580 –> 00:01:44,040 And there we have the Borel

    60 00:01:44,309 –> 00:01:45,370 Sigma algebra.

    61 00:01:46,650 –> 00:01:48,309 Also recall

    62 00:01:48,319 –> 00:01:50,309 that measurable means

    63 00:01:50,519 –> 00:01:52,209 that all the preimages

    64 00:01:52,220 –> 00:01:53,959 from elements from the sigma

    65 00:01:53,970 –> 00:01:55,910 algebra here lie

    66 00:01:55,919 –> 00:01:57,230 in the sigma algebra A.

    67 00:01:57,239 –> 00:01:59,110 In other

    68 00:01:59,120 –> 00:02:00,860 words, preimage

    69 00:02:00,870 –> 00:02:02,720 of a set

    70 00:02:02,730 –> 00:02:04,599 E is in

    71 00:02:04,610 –> 00:02:06,519 A for all

    72 00:02:06,529 –> 00:02:08,100 Borel sets E

    73 00:02:08,630 –> 00:02:09,139 OK.

    74 00:02:09,149 –> 00:02:10,839 So these are the functions

    75 00:02:10,850 –> 00:02:12,160 where we want to be able

    76 00:02:12,169 –> 00:02:13,940 to integrate them in the

    77 00:02:13,949 –> 00:02:14,320 end.

    78 00:02:15,160 –> 00:02:16,910 However, at the moment,

    79 00:02:17,169 –> 00:02:18,910 they might be too complicated.

    80 00:02:18,919 –> 00:02:20,470 So we start with functions

    81 00:02:20,479 –> 00:02:21,589 that we already know.

    82 00:02:22,339 –> 00:02:24,070 For example, we already

    83 00:02:24,080 –> 00:02:25,710 know that the characteristic

    84 00:02:25,720 –> 00:02:27,669 function is a measurable

    85 00:02:27,679 –> 00:02:28,309 map.

    86 00:02:28,320 –> 00:02:30,229 If we choose a measurable

    87 00:02:30,240 –> 00:02:30,669 set,

    88 00:02:32,050 –> 00:02:33,240 this means A

    89 00:02:33,529 –> 00:02:35,350 lies in the Sigma algebra.

    90 00:02:36,419 –> 00:02:38,020 We already know how to sketch

    91 00:02:38,029 –> 00:02:39,080 this function.

    92 00:02:39,880 –> 00:02:41,619 If we have our abstract X

    93 00:02:41,630 –> 00:02:43,179 here on the line

    94 00:02:43,330 –> 00:02:45,059 and maybe this

    95 00:02:45,070 –> 00:02:46,539 is the set A.

    96 00:02:46,809 –> 00:02:48,550 So these two things together

    97 00:02:48,559 –> 00:02:50,529 are A, then we

    98 00:02:50,539 –> 00:02:52,479 can sketch the graph of

    99 00:02:52,490 –> 00:02:54,029 our characteristic function.

    100 00:02:56,149 –> 00:02:57,820 It will be zero here

    101 00:02:58,119 –> 00:02:59,789 and has to value you

    102 00:02:59,800 –> 00:03:01,759 one where the set A

    103 00:03:01,770 –> 00:03:02,279 lies.

    104 00:03:02,289 –> 00:03:04,179 So here it will be one and

    105 00:03:04,190 –> 00:03:05,490 here also zero

    106 00:03:05,789 –> 00:03:07,720 and there also zero

    107 00:03:08,589 –> 00:03:10,160 for visualization of the

    108 00:03:10,169 –> 00:03:12,059 integral, it’s always

    109 00:03:12,070 –> 00:03:14,000 good to see the integral

    110 00:03:14,009 –> 00:03:15,970 as the area below

    111 00:03:15,979 –> 00:03:16,779 the graph

    112 00:03:16,789 –> 00:03:17,960 And the x axis

    113 00:03:18,820 –> 00:03:19,809 here, this would mean we

    114 00:03:19,820 –> 00:03:21,550 look at this area

    115 00:03:21,559 –> 00:03:22,190 here

    116 00:03:23,690 –> 00:03:25,190 and the area here.

    117 00:03:29,149 –> 00:03:30,639 And because the value of

    118 00:03:30,649 –> 00:03:32,289 the function is just one,

    119 00:03:32,300 –> 00:03:34,250 it does not matter how abstract

    120 00:03:34,259 –> 00:03:35,330 this whole measure space

    121 00:03:35,339 –> 00:03:37,130 is, this area should

    122 00:03:37,139 –> 00:03:39,130 be exactly the same as

    123 00:03:39,139 –> 00:03:41,009 the volume, the measure

    124 00:03:41,020 –> 00:03:42,279 of the set A

    125 00:03:43,160 –> 00:03:44,380 or in other words, a

    126 00:03:44,389 –> 00:03:46,059 meaningful integral

    127 00:03:46,070 –> 00:03:46,740 notion.

    128 00:03:46,750 –> 00:03:48,179 And maybe let’s call it just

    129 00:03:48,190 –> 00:03:49,330 by capital I

    130 00:03:49,889 –> 00:03:51,369 of this characteristic

    131 00:03:51,380 –> 00:03:52,899 function should always

    132 00:03:52,910 –> 00:03:54,460 fulfill that.

    133 00:03:54,470 –> 00:03:56,039 This integral

    134 00:03:56,169 –> 00:03:58,029 is the measure of the

    135 00:03:58,039 –> 00:03:58,949 set A.

    136 00:04:00,089 –> 00:04:01,470 Now we have a new symbol

    137 00:04:01,479 –> 00:04:03,020 on the left I of a

    138 00:04:03,029 –> 00:04:03,800 function.

    139 00:04:03,809 –> 00:04:05,750 So you should read this indeed,

    140 00:04:05,770 –> 00:04:07,149 as a definition,

    141 00:04:08,520 –> 00:04:10,270 this means now you know

    142 00:04:10,279 –> 00:04:12,210 how to integrate characteristic

    143 00:04:12,220 –> 00:04:12,740 functions.

    144 00:04:13,949 –> 00:04:15,050 In fact, there are other

    145 00:04:15,059 –> 00:04:16,329 functions where we can

    146 00:04:16,339 –> 00:04:18,079 define the integral

    147 00:04:18,089 –> 00:04:19,798 also in such a simple

    148 00:04:19,809 –> 00:04:20,190 way.

    149 00:04:20,380 –> 00:04:22,269 And these are called simple

    150 00:04:22,279 –> 00:04:23,089 functions.

    151 00:04:23,950 –> 00:04:25,649 There are also a lot of other

    152 00:04:25,660 –> 00:04:27,570 names that one uses for such

    153 00:04:27,579 –> 00:04:28,649 simple functions.

    154 00:04:29,559 –> 00:04:30,940 For example, step

    155 00:04:30,950 –> 00:04:32,470 functions as I

    156 00:04:32,480 –> 00:04:34,029 did in the

    157 00:04:34,040 –> 00:04:35,640 title of this video

    158 00:04:36,059 –> 00:04:36,940 and also

    159 00:04:37,000 –> 00:04:38,859 staircase functions

    160 00:04:39,649 –> 00:04:41,100 and also other

    161 00:04:41,109 –> 00:04:41,839 names.

    162 00:04:42,779 –> 00:04:44,489 And you see the name

    163 00:04:44,500 –> 00:04:46,359 might not be important.

    164 00:04:46,380 –> 00:04:47,809 You just have to know what

    165 00:04:47,820 –> 00:04:49,730 such simple functions are.

    166 00:04:50,799 –> 00:04:51,429 In short.

    167 00:04:51,440 –> 00:04:53,220 A simple function F

    168 00:04:53,230 –> 00:04:54,730 is just a linear

    169 00:04:54,739 –> 00:04:56,660 combination of such

    170 00:04:56,670 –> 00:04:58,140 characteristic functions.

    171 00:04:58,660 –> 00:05:00,130 This means you can write

    172 00:05:00,140 –> 00:05:01,890 F of X as a

    173 00:05:01,899 –> 00:05:02,739 sum.

    174 00:05:02,869 –> 00:05:04,170 Where we will start with

    175 00:05:04,200 –> 00:05:06,070 one and go to a fixed

    176 00:05:06,079 –> 00:05:07,820 N and where we

    177 00:05:07,829 –> 00:05:09,679 have numbers C

    178 00:05:09,690 –> 00:05:11,149 I and

    179 00:05:11,160 –> 00:05:12,649 characteristic functions

    180 00:05:12,660 –> 00:05:14,250 corresponding to some

    181 00:05:14,260 –> 00:05:16,209 measurable set A

    182 00:05:16,220 –> 00:05:16,850 I.

    183 00:05:18,619 –> 00:05:19,809 In other words, a

    184 00:05:19,829 –> 00:05:21,779 function F is called

    185 00:05:21,790 –> 00:05:22,559 simple.

    186 00:05:22,570 –> 00:05:23,910 If you can find

    187 00:05:23,920 –> 00:05:25,809 measurable sets A_1

    188 00:05:25,820 –> 00:05:26,869 till A_n

    189 00:05:27,750 –> 00:05:29,390 and real numbers c_1

    190 00:05:29,399 –> 00:05:31,200 to c_n

    191 00:05:31,380 –> 00:05:32,839 such that you can write the

    192 00:05:32,850 –> 00:05:34,600 function in this form

    193 00:05:35,500 –> 00:05:36,899 because we know that the

    194 00:05:36,910 –> 00:05:38,260 characteristic functions

    195 00:05:38,269 –> 00:05:40,100 are measurable and also

    196 00:05:40,109 –> 00:05:42,000 sums of measurable

    197 00:05:42,010 –> 00:05:43,959 maps are also measurable.

    198 00:05:44,059 –> 00:05:45,910 We also know that this

    199 00:05:45,920 –> 00:05:47,160 simple function is

    200 00:05:47,170 –> 00:05:48,000 therefore

    201 00:05:48,019 –> 00:05:49,239 measurable.

    202 00:05:50,519 –> 00:05:52,040 Also here, I would

    203 00:05:52,049 –> 00:05:53,700 say to get some

    204 00:05:53,709 –> 00:05:55,500 visualization, let’s

    205 00:05:55,510 –> 00:05:57,049 sketch the graph of such

    206 00:05:57,059 –> 00:05:57,899 a simple function.

    207 00:05:58,980 –> 00:06:00,709 This is the same as before.

    208 00:06:00,720 –> 00:06:02,619 So we have X here and

    209 00:06:02,630 –> 00:06:04,540 R here and

    210 00:06:04,549 –> 00:06:06,540 maybe here now we have set

    211 00:06:06,549 –> 00:06:08,450 A_1 and

    212 00:06:08,459 –> 00:06:09,769 here a set A_2

    213 00:06:09,779 –> 00:06:10,470 .

    214 00:06:10,899 –> 00:06:12,160 And let’s choose a split

    215 00:06:12,170 –> 00:06:13,869 set A three here.

    216 00:06:15,989 –> 00:06:17,980 And for these sets, we can

    217 00:06:17,989 –> 00:06:19,589 now choose some

    218 00:06:19,600 –> 00:06:21,100 constant c_i.

    219 00:06:21,869 –> 00:06:23,100 Hence, we have for the graph

    220 00:06:23,109 –> 00:06:24,869 of the function some value

    221 00:06:24,880 –> 00:06:26,609 here over A_1,

    222 00:06:26,619 –> 00:06:28,220 another value over A_2

    223 00:06:28,230 –> 00:06:29,209 here.

    224 00:06:29,640 –> 00:06:31,239 And maybe a lower one

    225 00:06:31,250 –> 00:06:33,010 here for A_3

    226 00:06:34,890 –> 00:06:36,820 and outside of this set, we

    227 00:06:36,829 –> 00:06:38,089 are just at zero.

    228 00:06:38,920 –> 00:06:40,420 If we now go back to our

    229 00:06:40,429 –> 00:06:41,760 visualization of the

    230 00:06:41,769 –> 00:06:43,420 integral, we know it’s given

    231 00:06:43,429 –> 00:06:44,459 by rectangles.

    232 00:06:44,470 –> 00:06:46,320 For example, here we have

    233 00:06:46,329 –> 00:06:47,320 the rectangle

    234 00:06:49,739 –> 00:06:51,549 with length given by a

    235 00:06:51,559 –> 00:06:53,239 one and the height

    236 00:06:53,250 –> 00:06:54,640 given by C one.

    237 00:06:55,619 –> 00:06:57,420 Therefore, this rectangle

    238 00:06:57,429 –> 00:06:58,549 here just

    239 00:06:58,559 –> 00:07:00,160 represents the part of the

    240 00:07:00,170 –> 00:07:01,959 integral that is given

    241 00:07:01,970 –> 00:07:03,799 by C one times the

    242 00:07:03,809 –> 00:07:05,540 measure of a one.

    243 00:07:06,959 –> 00:07:08,820 And the contribution for

    244 00:07:08,829 –> 00:07:10,359 the integral given by the

    245 00:07:10,369 –> 00:07:12,100 set A two is just

    246 00:07:12,109 –> 00:07:13,500 this rectangle here.

    247 00:07:14,549 –> 00:07:16,130 And of course, now given

    248 00:07:16,140 –> 00:07:17,609 by C two

    249 00:07:17,649 –> 00:07:19,329 times the measure

    250 00:07:19,339 –> 00:07:20,570 of A two.

    251 00:07:21,269 –> 00:07:22,489 And then the last part is

    252 00:07:22,500 –> 00:07:24,410 coming from a three which

    253 00:07:24,420 –> 00:07:25,790 consists of three parts.

    254 00:07:25,799 –> 00:07:27,779 But in the sum, of course,

    255 00:07:27,790 –> 00:07:29,230 this is also C

    256 00:07:29,239 –> 00:07:30,660 three times the

    257 00:07:30,670 –> 00:07:32,149 measure of a

    258 00:07:32,160 –> 00:07:32,709 three.

    259 00:07:33,630 –> 00:07:35,010 And now you should see that

    260 00:07:35,019 –> 00:07:36,790 we just have to add

    261 00:07:36,799 –> 00:07:37,690 up all these

    262 00:07:37,700 –> 00:07:39,500 contributions to get the

    263 00:07:39,510 –> 00:07:40,929 full integral, to get the

    264 00:07:40,940 –> 00:07:42,890 full area in this picture.

    265 00:07:44,000 –> 00:07:45,540 Hence, again, for

    266 00:07:45,549 –> 00:07:46,920 meaningful integral

    267 00:07:46,929 –> 00:07:48,899 definition, we need a sum

    268 00:07:48,910 –> 00:07:50,149 of all these parts.

    269 00:07:50,160 –> 00:07:51,859 So I of

    270 00:07:51,869 –> 00:07:53,700 F is then

    271 00:07:53,709 –> 00:07:55,700 given by a sum

    272 00:07:56,929 –> 00:07:58,869 and they have the same C

    273 00:07:58,880 –> 00:08:00,549 I as before, but

    274 00:08:00,559 –> 00:08:02,549 now not the characteristic

    275 00:08:02,559 –> 00:08:03,589 function but the

    276 00:08:03,600 –> 00:08:05,480 corresponding measures

    277 00:08:05,640 –> 00:08:06,750 of the sets.

    278 00:08:08,609 –> 00:08:10,510 And indeed, this will be

    279 00:08:10,519 –> 00:08:12,380 our integral for

    280 00:08:12,390 –> 00:08:14,070 simple functions in the end.

    281 00:08:15,660 –> 00:08:17,399 However, at the moment,

    282 00:08:17,519 –> 00:08:18,950 you should see immediately

    283 00:08:18,959 –> 00:08:20,929 one problem with this definition.

    284 00:08:22,059 –> 00:08:24,049 What happens if we have

    285 00:08:24,059 –> 00:08:25,480 a set with volume

    286 00:08:25,489 –> 00:08:26,480 infinity?

    287 00:08:26,489 –> 00:08:28,040 So that this number

    288 00:08:28,049 –> 00:08:29,640 here is the symbol

    289 00:08:29,649 –> 00:08:30,559 infinity.

    290 00:08:31,450 –> 00:08:33,429 There we could have for example,

    291 00:08:34,090 –> 00:08:35,469 three as the C

    292 00:08:35,760 –> 00:08:37,349 times infinity.

    293 00:08:38,330 –> 00:08:39,549 Maybe that’s not a problem

    294 00:08:39,558 –> 00:08:40,909 for you because it’s just

    295 00:08:40,919 –> 00:08:42,460 infinity itself again.

    296 00:08:42,739 –> 00:08:44,179 But then what happens if

    297 00:08:44,190 –> 00:08:46,109 the next C here is

    298 00:08:46,119 –> 00:08:48,070 minus two and it measures

    299 00:08:48,080 –> 00:08:49,150 also infinity.

    300 00:08:50,590 –> 00:08:51,979 And of course, this is not

    301 00:08:51,989 –> 00:08:53,890 simply just one infinity.

    302 00:08:53,900 –> 00:08:55,260 This is not defined at the

    303 00:08:55,270 –> 00:08:55,770 moment.

    304 00:08:57,070 –> 00:08:58,469 In fact, we have a

    305 00:08:58,479 –> 00:09:00,229 problem with the definition

    306 00:09:00,239 –> 00:09:00,859 in general.

    307 00:09:02,169 –> 00:09:04,140 In order to solve this problem,

    308 00:09:04,239 –> 00:09:05,539 there are essentially two

    309 00:09:05,549 –> 00:09:06,770 ways one can go.

    310 00:09:06,780 –> 00:09:08,510 Now on the one

    311 00:09:08,520 –> 00:09:10,179 hand, one can restrict

    312 00:09:10,190 –> 00:09:12,039 oneself to such simple

    313 00:09:12,049 –> 00:09:13,979 functions where the corresponding

    314 00:09:14,140 –> 00:09:16,020 AIS only have

    315 00:09:16,030 –> 00:09:17,200 finite measure.

    316 00:09:18,010 –> 00:09:19,559 Obviously, then there’s no

    317 00:09:19,570 –> 00:09:21,429 problem with such infinities

    318 00:09:21,440 –> 00:09:22,659 because there is no simple

    319 00:09:22,669 –> 00:09:24,020 infinity involved in the

    320 00:09:24,030 –> 00:09:24,750 whole sum.

    321 00:09:25,570 –> 00:09:27,049 And on the other hand, one

    322 00:09:27,059 –> 00:09:28,789 can demand that these

    323 00:09:28,799 –> 00:09:30,179 constants have to be

    324 00:09:30,190 –> 00:09:32,119 positive, then there

    325 00:09:32,130 –> 00:09:33,650 might be infinity in the

    326 00:09:33,659 –> 00:09:35,609 sun, but only with the same

    327 00:09:35,619 –> 00:09:36,250 sign.

    328 00:09:36,960 –> 00:09:38,280 And then of course, we don’t

    329 00:09:38,289 –> 00:09:39,830 have a problem adding up

    330 00:09:39,840 –> 00:09:41,380 positive infinities.

    331 00:09:42,109 –> 00:09:44,030 And exactly this second

    332 00:09:44,039 –> 00:09:45,890 possibility is the way

    333 00:09:45,900 –> 00:09:47,159 I want to go here.

    334 00:09:48,059 –> 00:09:49,960 Now, finally, we can put

    335 00:09:49,969 –> 00:09:51,929 the integral into a

    336 00:09:51,940 –> 00:09:52,799 definition

    337 00:09:53,890 –> 00:09:54,520 for this.

    338 00:09:54,530 –> 00:09:56,440 We will consider now, just

    339 00:09:56,450 –> 00:09:58,299 functions defined on X.

    340 00:09:59,070 –> 00:10:00,809 And as before, there should

    341 00:10:00,820 –> 00:10:02,719 be step or simple

    342 00:10:02,729 –> 00:10:03,409 functions.

    343 00:10:04,140 –> 00:10:05,880 Moreover, we want now to

    344 00:10:05,890 –> 00:10:07,859 consider only the positive

    345 00:10:07,869 –> 00:10:09,659 ones or more concretely

    346 00:10:09,669 –> 00:10:11,460 the non-negative ones.

    347 00:10:12,599 –> 00:10:13,979 The whole set of such

    348 00:10:13,989 –> 00:10:15,380 functions is now

    349 00:10:15,390 –> 00:10:17,150 denoted by a

    350 00:10:17,159 –> 00:10:19,059 curved S where I put

    351 00:10:19,070 –> 00:10:20,460 in A plus

    352 00:10:21,789 –> 00:10:23,609 and this S plus is now

    353 00:10:23,619 –> 00:10:25,609 almost a vector space

    354 00:10:25,950 –> 00:10:27,559 in the sense that we can

    355 00:10:27,570 –> 00:10:29,090 add functions as we want.

    356 00:10:29,349 –> 00:10:30,909 But we can’t scale them as

    357 00:10:30,919 –> 00:10:32,369 we want because

    358 00:10:32,380 –> 00:10:34,309 only positive scalars are

    359 00:10:34,320 –> 00:10:35,799 allowed otherwise we would

    360 00:10:35,809 –> 00:10:36,780 leave this space.

    361 00:10:36,789 –> 00:10:38,570 So you should imagine

    362 00:10:38,580 –> 00:10:40,169 this as a half space.

    363 00:10:41,460 –> 00:10:43,179 If we go now back to our

    364 00:10:43,190 –> 00:10:44,719 definition of simple

    365 00:10:44,729 –> 00:10:45,859 functions, you

    366 00:10:45,869 –> 00:10:47,599 immediately recognize

    367 00:10:47,760 –> 00:10:49,239 that this representation

    368 00:10:49,250 –> 00:10:50,880 here is not unique for the

    369 00:10:50,890 –> 00:10:51,599 function F.

    370 00:10:52,210 –> 00:10:54,030 For example, we could split

    371 00:10:54,039 –> 00:10:55,760 the set A two into two

    372 00:10:55,770 –> 00:10:56,429 parts.

    373 00:10:56,659 –> 00:10:58,140 And then we have one

    374 00:10:58,150 –> 00:11:00,119 term in addition in

    375 00:11:00,130 –> 00:11:02,080 the sum here, but

    376 00:11:02,090 –> 00:11:03,760 it does not change the function

    377 00:11:03,770 –> 00:11:04,700 F itself.

    378 00:11:04,710 –> 00:11:06,119 So the graph looks still

    379 00:11:06,130 –> 00:11:07,799 the same just this

    380 00:11:07,809 –> 00:11:09,119 sum is different.

    381 00:11:10,030 –> 00:11:11,809 Therefore, another description

    382 00:11:11,820 –> 00:11:13,710 which is independent of this

    383 00:11:13,719 –> 00:11:15,309 representation for a simple

    384 00:11:15,320 –> 00:11:16,809 function would be

    385 00:11:18,229 –> 00:11:19,700 I have a measurable

    386 00:11:19,710 –> 00:11:20,369 function.

    387 00:11:20,440 –> 00:11:21,739 And it only takes

    388 00:11:21,750 –> 00:11:23,700 finitely many values.

    389 00:11:24,650 –> 00:11:26,549 In this case, I can always

    390 00:11:26,559 –> 00:11:28,309 find such a representation

    391 00:11:28,320 –> 00:11:28,630 here.

    392 00:11:29,799 –> 00:11:30,780 The idea is now

    393 00:11:30,789 –> 00:11:32,400 always choose a

    394 00:11:32,409 –> 00:11:33,380 suitable

    395 00:11:33,390 –> 00:11:34,969 representation for your

    396 00:11:34,979 –> 00:11:36,140 simple function F.

    397 00:11:36,950 –> 00:11:38,770 In our case here, if

    398 00:11:38,780 –> 00:11:40,609 F comes from our S

    399 00:11:40,619 –> 00:11:41,440 plus,

    400 00:11:43,619 –> 00:11:45,479 then choose a representation

    401 00:11:45,659 –> 00:11:46,640 as always.

    402 00:11:46,650 –> 00:11:48,520 So we can write it as a sum.

    403 00:11:49,340 –> 00:11:51,289 But now where the c_i’s

    404 00:11:51,299 –> 00:11:52,770 are also non-

    405 00:11:52,799 –> 00:11:53,559 negative.

    406 00:11:55,109 –> 00:11:56,580 Of course, if the function

    407 00:11:56,590 –> 00:11:58,049 is non-negative, then we

    408 00:11:58,059 –> 00:11:59,669 can always choose such a

    409 00:11:59,679 –> 00:12:00,570 representation.

    410 00:12:01,880 –> 00:12:03,679 And now it’s very easy

    411 00:12:03,690 –> 00:12:05,419 to define the Lebesgue integral.

    412 00:12:06,429 –> 00:12:08,320 So this is our new notion

    413 00:12:08,780 –> 00:12:10,570 and of course, often we will

    414 00:12:10,580 –> 00:12:12,419 ignore the Lebesgue and

    415 00:12:12,429 –> 00:12:13,710 just speak of the

    416 00:12:13,719 –> 00:12:15,530 integral because this is

    417 00:12:15,539 –> 00:12:17,530 the integral we define here.

    418 00:12:18,419 –> 00:12:19,789 And to be more concrete,

    419 00:12:19,799 –> 00:12:21,190 we also would say with

    420 00:12:21,200 –> 00:12:22,799 respect to our

    421 00:12:22,809 –> 00:12:24,409 given measure mu.

    422 00:12:25,809 –> 00:12:27,349 And this integral is now

    423 00:12:27,359 –> 00:12:29,070 given as we already

    424 00:12:29,080 –> 00:12:29,510 know.

    425 00:12:29,520 –> 00:12:31,390 And also denoted by

    426 00:12:31,400 –> 00:12:33,340 I of F where we

    427 00:12:33,349 –> 00:12:34,859 just add up

    428 00:12:35,150 –> 00:12:36,780 the measures of the

    429 00:12:36,789 –> 00:12:38,539 sets A I where we

    430 00:12:38,549 –> 00:12:40,530 scale them also by the

    431 00:12:40,539 –> 00:12:41,169 CIS.

    432 00:12:42,039 –> 00:12:43,260 So we have C I

    433 00:12:43,270 –> 00:12:44,900 times the measure

    434 00:12:44,909 –> 00:12:46,369 of A I.

    435 00:12:47,710 –> 00:12:49,590 And because the C I are

    436 00:12:49,599 –> 00:12:51,409 non-negative and the measure

    437 00:12:51,419 –> 00:12:53,090 itself is non-negative or

    438 00:12:53,099 –> 00:12:54,260 in the worst case just

    439 00:12:54,270 –> 00:12:56,030 infinity, we know this

    440 00:12:56,039 –> 00:12:58,020 integral is now also in the

    441 00:12:58,070 –> 00:12:59,729 interval zero

    442 00:12:59,840 –> 00:13:01,390 to infinity where we

    443 00:13:01,400 –> 00:13:03,090 include both parts.

    444 00:13:04,219 –> 00:13:05,479 One fact I can give you

    445 00:13:05,489 –> 00:13:07,299 immediately this definition

    446 00:13:07,309 –> 00:13:08,929 for integral is a

    447 00:13:08,940 –> 00:13:10,400 well-defined

    448 00:13:10,510 –> 00:13:11,400 object.

    449 00:13:13,080 –> 00:13:14,940 This means that it does not

    450 00:13:14,950 –> 00:13:16,520 depend on the chosen

    451 00:13:16,530 –> 00:13:18,320 representation for our simple

    452 00:13:18,330 –> 00:13:19,380 function F.

    453 00:13:19,619 –> 00:13:20,859 So if you choose another

    454 00:13:20,869 –> 00:13:22,260 representation with this

    455 00:13:22,270 –> 00:13:24,260 property, you get the same

    456 00:13:24,270 –> 00:13:26,099 number out, maybe

    457 00:13:26,109 –> 00:13:27,489 this is a thing you could

    458 00:13:27,500 –> 00:13:28,580 try to prove.

    459 00:13:29,409 –> 00:13:31,070 However, of course, the

    460 00:13:31,080 –> 00:13:32,369 visualization should be the

    461 00:13:32,380 –> 00:13:33,369 important thing.

    462 00:13:33,409 –> 00:13:35,280 So it does not matter how

    463 00:13:35,289 –> 00:13:36,750 you split up these

    464 00:13:36,760 –> 00:13:37,919 rectangles here.

    465 00:13:38,090 –> 00:13:39,710 Yeah, the sum or the sum

    466 00:13:39,719 –> 00:13:41,630 of the areas should be always

    467 00:13:41,640 –> 00:13:43,330 the same no

    468 00:13:43,340 –> 00:13:44,929 matter how you split up the

    469 00:13:44,940 –> 00:13:46,809 set here on the X axis or

    470 00:13:46,820 –> 00:13:48,659 you split up the C

    471 00:13:48,669 –> 00:13:50,359 values on the Y axis.

    472 00:13:51,349 –> 00:13:52,700 In other words, it makes

    473 00:13:52,710 –> 00:13:54,270 sense that this integral

    474 00:13:54,280 –> 00:13:55,760 here is indeed well-

    475 00:13:55,770 –> 00:13:56,369 defined.

    476 00:13:57,380 –> 00:13:59,359 And finally, I can give you

    477 00:13:59,369 –> 00:14:01,099 the usual notation one

    478 00:14:01,109 –> 00:14:02,419 uses for the integral,

    479 00:14:03,070 –> 00:14:04,900 of course, an integral sign,

    480 00:14:04,919 –> 00:14:06,640 we put the set X

    481 00:14:06,650 –> 00:14:08,280 here then comes the

    482 00:14:08,289 –> 00:14:10,010 function F and

    483 00:14:10,020 –> 00:14:11,530 then the measure mu

    484 00:14:11,539 –> 00:14:13,380 by d

    485 00:14:13,539 –> 00:14:13,929 mu

    486 00:14:15,130 –> 00:14:16,530 sometimes also a

    487 00:14:16,539 –> 00:14:17,950 variable is needed.

    488 00:14:17,960 –> 00:14:19,650 And the notation looks almost

    489 00:14:19,659 –> 00:14:20,349 the same.

    490 00:14:20,859 –> 00:14:21,830 So you just

    491 00:14:21,950 –> 00:14:23,549 include a variable

    492 00:14:23,559 –> 00:14:24,950 name for the function.

    493 00:14:24,960 –> 00:14:26,440 So a lower case X

    494 00:14:26,450 –> 00:14:28,030 here and then d mu

    495 00:14:28,840 –> 00:14:29,619 X.

    496 00:14:31,030 –> 00:14:32,890 So this is the Lebesgue integral

    497 00:14:32,929 –> 00:14:34,729 for step or

    498 00:14:34,739 –> 00:14:35,890 simple functions.

    499 00:14:36,640 –> 00:14:38,219 The idea is now that we

    500 00:14:38,229 –> 00:14:40,059 expand this definition

    501 00:14:40,070 –> 00:14:41,489 such that we can also

    502 00:14:41,500 –> 00:14:43,159 integrate more complicated

    503 00:14:43,169 –> 00:14:43,809 functions.

    504 00:14:44,659 –> 00:14:46,340 But before we do that, let’s

    505 00:14:46,349 –> 00:14:48,320 first look at some properties

    506 00:14:48,330 –> 00:14:49,719 this integral has

    507 00:14:50,419 –> 00:14:51,919 firstly, it is

    508 00:14:51,929 –> 00:14:53,489 almost linear.

    509 00:14:53,960 –> 00:14:55,590 I already told you

    510 00:14:55,619 –> 00:14:57,130 we almost have vector

    511 00:14:57,140 –> 00:14:59,030 space for our S plus.

    512 00:14:59,179 –> 00:15:00,630 And in this sense, it’s

    513 00:15:00,640 –> 00:15:02,450 linear, we

    514 00:15:02,460 –> 00:15:03,900 just have to restrict the

    515 00:15:03,909 –> 00:15:05,890 scalars to non-negative

    516 00:15:05,900 –> 00:15:06,679 numbers.

    517 00:15:07,440 –> 00:15:08,780 And then we can pull out

    518 00:15:08,789 –> 00:15:09,900 the addition sign.

    519 00:15:09,909 –> 00:15:11,809 And also the scalars,

    520 00:15:12,309 –> 00:15:14,250 hence this equality

    521 00:15:14,260 –> 00:15:15,809 holds for all simple

    522 00:15:15,820 –> 00:15:17,640 functions FG and

    523 00:15:17,650 –> 00:15:19,010 for all scalars

    524 00:15:19,179 –> 00:15:21,070 alpha beta greater

    525 00:15:21,080 –> 00:15:22,299 or equal to zero.

    526 00:15:23,359 –> 00:15:25,010 And of course, this immediately

    527 00:15:25,020 –> 00:15:26,820 comes from this sum here.

    528 00:15:28,119 –> 00:15:29,979 Also by using a, we

    529 00:15:29,989 –> 00:15:31,780 can now show that if we

    530 00:15:31,789 –> 00:15:33,070 have a step function or a

    531 00:15:33,080 –> 00:15:34,940 simple function that is always

    532 00:15:34,950 –> 00:15:36,219 bigger than another one,

    533 00:15:36,229 –> 00:15:37,989 so we have F less or

    534 00:15:38,000 –> 00:15:39,140 equal than G,

    535 00:15:39,799 –> 00:15:41,419 then also the area

    536 00:15:41,530 –> 00:15:43,280 between the graph

    537 00:15:43,289 –> 00:15:44,780 and the X axis should be

    538 00:15:44,789 –> 00:15:46,619 bigger for G than for F.

    539 00:15:47,150 –> 00:15:48,349 And of course, this is what

    540 00:15:48,359 –> 00:15:49,359 the integral tells us.

    541 00:15:49,369 –> 00:15:51,280 So I of F

    542 00:15:51,289 –> 00:15:53,179 is also less or equal

    543 00:15:53,190 –> 00:15:54,960 than I of G.

    544 00:15:56,020 –> 00:15:57,599 And this is what one calls

    545 00:15:57,609 –> 00:15:58,710 monotonicity.

    546 00:15:59,659 –> 00:16:01,270 In fact, this monotonic

    547 00:16:01,280 –> 00:16:03,109 behavior is what we now

    548 00:16:03,119 –> 00:16:04,989 can use to expand our

    549 00:16:05,000 –> 00:16:06,729 definition to

    550 00:16:06,739 –> 00:16:08,729 general measurable maps

    551 00:16:09,390 –> 00:16:10,250 in order to get a

    552 00:16:10,260 –> 00:16:12,039 visualization, let’s

    553 00:16:12,049 –> 00:16:13,919 sketch again a graph.

    554 00:16:14,739 –> 00:16:16,619 Now this is the graph of

    555 00:16:16,630 –> 00:16:18,119 some measurable map

    556 00:16:18,349 –> 00:16:20,049 which is not a simple function.

    557 00:16:20,059 –> 00:16:20,840 As you can see.

    558 00:16:21,719 –> 00:16:23,299 Now, the integral should

    559 00:16:23,309 –> 00:16:25,130 be now again represented

    560 00:16:25,140 –> 00:16:26,940 by the area below

    561 00:16:26,950 –> 00:16:28,880 the graph and the X axis.

    562 00:16:29,799 –> 00:16:31,429 However, we only have this

    563 00:16:31,440 –> 00:16:33,190 notion for the moment for

    564 00:16:33,200 –> 00:16:34,469 the simple functions.

    565 00:16:35,270 –> 00:16:36,390 And as you know, by now,

    566 00:16:36,400 –> 00:16:38,090 a simple function has

    567 00:16:38,099 –> 00:16:39,469 only finitely many

    568 00:16:39,479 –> 00:16:40,260 values.

    569 00:16:40,679 –> 00:16:41,770 And this blue curve here

    570 00:16:41,780 –> 00:16:43,650 shows you there are infinitely

    571 00:16:43,659 –> 00:16:44,770 many values.

    572 00:16:45,539 –> 00:16:46,929 The whole idea of the Lebesgue

    573 00:16:47,080 –> 00:16:48,770 integration is now to

    574 00:16:48,780 –> 00:16:50,440 approximate the function

    575 00:16:50,450 –> 00:16:51,830 by finitely many

    576 00:16:51,840 –> 00:16:52,469 values.

    577 00:16:53,549 –> 00:16:55,530 So I just choose some values

    578 00:16:55,539 –> 00:16:57,010 here on the y axis

    579 00:16:57,020 –> 00:16:58,549 such that the whole

    580 00:16:58,869 –> 00:17:00,320 curve is in some sense

    581 00:17:00,330 –> 00:17:01,049 covered.

    582 00:17:02,390 –> 00:17:04,020 And now the idea is to

    583 00:17:04,030 –> 00:17:05,800 define a suitable

    584 00:17:05,810 –> 00:17:06,848 simple function.

    585 00:17:08,218 –> 00:17:09,608 Therefore, maybe I choose

    586 00:17:09,618 –> 00:17:11,159 this interval here

    587 00:17:11,568 –> 00:17:13,358 and look what happens to

    588 00:17:13,368 –> 00:17:14,968 the values of the function

    589 00:17:14,979 –> 00:17:15,588 there.

    590 00:17:15,878 –> 00:17:17,458 And now you see we have two

    591 00:17:17,468 –> 00:17:19,358 parts that are

    592 00:17:19,368 –> 00:17:21,138 mapped to this

    593 00:17:21,148 –> 00:17:22,038 interval here.

    594 00:17:22,939 –> 00:17:24,550 For the X axis, this

    595 00:17:24,560 –> 00:17:25,489 looks like this.

    596 00:17:25,500 –> 00:17:27,390 So we have this part here

    597 00:17:27,400 –> 00:17:29,209 that ends here

    598 00:17:29,520 –> 00:17:31,290 and we have this part that

    599 00:17:31,300 –> 00:17:32,560 starts here

    600 00:17:32,949 –> 00:17:34,780 and also ends

    601 00:17:34,790 –> 00:17:35,209 here.

    602 00:17:36,109 –> 00:17:38,050 This means we have one part

    603 00:17:38,060 –> 00:17:39,979 of our set here and the other

    604 00:17:39,989 –> 00:17:41,530 part here.

    605 00:17:42,810 –> 00:17:44,170 And of course, this should

    606 00:17:44,180 –> 00:17:45,800 be our set that we

    607 00:17:45,810 –> 00:17:47,430 usually call A

    608 00:17:47,619 –> 00:17:48,270 I.

    609 00:17:48,280 –> 00:17:49,729 So this is just one A

    610 00:17:49,739 –> 00:17:50,810 I and the

    611 00:17:50,819 –> 00:17:52,479 corresponding C

    612 00:17:52,619 –> 00:17:54,349 you find now here.

    613 00:17:54,500 –> 00:17:56,459 So this would be our C I.

    614 00:17:57,859 –> 00:17:59,750 So I chose the lower part

    615 00:17:59,760 –> 00:18:01,199 of the interval because then

    616 00:18:01,209 –> 00:18:03,199 our step function is also

    617 00:18:03,260 –> 00:18:05,239 below the graph of the

    618 00:18:05,250 –> 00:18:05,839 function.

    619 00:18:05,849 –> 00:18:07,780 So this would be the

    620 00:18:07,790 –> 00:18:09,040 value of our step function

    621 00:18:09,050 –> 00:18:09,319 here.

    622 00:18:10,410 –> 00:18:12,339 And now you see all the other

    623 00:18:12,349 –> 00:18:14,040 values on the Y axis

    624 00:18:14,050 –> 00:18:15,810 give us the other

    625 00:18:15,819 –> 00:18:17,550 c_i’s and also the

    626 00:18:17,560 –> 00:18:19,180 corresponding A_i’s

    627 00:18:20,060 –> 00:18:21,880 therefore, with this decomposition

    628 00:18:21,890 –> 00:18:23,640 of the Y axis, we get

    629 00:18:23,650 –> 00:18:25,540 out the usual step function

    630 00:18:25,550 –> 00:18:27,489 where we have our C is there.

    631 00:18:27,569 –> 00:18:29,359 And the so defined A

    632 00:18:29,369 –> 00:18:31,300 is where we have now the

    633 00:18:31,310 –> 00:18:32,800 characteristic function here.

    634 00:18:33,670 –> 00:18:35,359 So this is a new simple function

    635 00:18:35,369 –> 00:18:36,630 which I should call

    636 00:18:36,859 –> 00:18:37,420 h now.

    637 00:18:38,310 –> 00:18:39,880 And the important part is

    638 00:18:39,890 –> 00:18:41,729 that h itself

    639 00:18:41,739 –> 00:18:43,010 lies always

    640 00:18:43,020 –> 00:18:44,989 below F so

    641 00:18:45,000 –> 00:18:46,609 always below the graph that

    642 00:18:46,619 –> 00:18:47,479 is blue here.

    643 00:18:48,729 –> 00:18:50,229 Therefore, if we want to

    644 00:18:50,239 –> 00:18:52,209 contain the monotonicity

    645 00:18:52,219 –> 00:18:53,619 in the general integral,

    646 00:18:53,630 –> 00:18:55,390 we now have an estimate

    647 00:18:55,400 –> 00:18:57,329 for the real integral value

    648 00:18:57,339 –> 00:18:58,989 of the function F.

    649 00:18:59,010 –> 00:19:00,160 We know that the integral

    650 00:19:00,170 –> 00:19:01,709 of this step function is

    651 00:19:01,719 –> 00:19:03,619 smaller than the real

    652 00:19:03,630 –> 00:19:04,349 integral.

    653 00:19:05,160 –> 00:19:07,020 And this gives now rise to

    654 00:19:07,030 –> 00:19:08,380 the following definition.

    655 00:19:09,319 –> 00:19:11,150 Hence, we choose a positive

    656 00:19:11,160 –> 00:19:13,140 or better a non-negative

    657 00:19:13,150 –> 00:19:14,500 function F that is

    658 00:19:14,510 –> 00:19:16,099 defined on our measure

    659 00:19:16,109 –> 00:19:17,060 space X.

    660 00:19:17,780 –> 00:19:18,959 And of course, it should

    661 00:19:18,969 –> 00:19:20,359 be measurable.

    662 00:19:21,420 –> 00:19:23,260 And now for each decomposition

    663 00:19:23,270 –> 00:19:25,060 of the Y axis, we can

    664 00:19:25,069 –> 00:19:26,260 choose such a

    665 00:19:26,270 –> 00:19:28,020 step function or simple

    666 00:19:28,030 –> 00:19:29,920 function h that lies

    667 00:19:29,930 –> 00:19:31,800 pointwisely below

    668 00:19:31,810 –> 00:19:32,930 the graph of the function

    669 00:19:32,939 –> 00:19:33,829 F itself.

    670 00:19:34,650 –> 00:19:36,500 And if we choose the function

    671 00:19:36,510 –> 00:19:38,369 h out of the

    672 00:19:38,510 –> 00:19:39,760 step function that are

    673 00:19:39,770 –> 00:19:41,660 positive or non negative.

    674 00:19:41,670 –> 00:19:43,589 So S plus, we know

    675 00:19:43,599 –> 00:19:44,790 we can look at the

    676 00:19:44,800 –> 00:19:46,500 integral of this

    677 00:19:46,510 –> 00:19:47,030 h.

    678 00:19:48,040 –> 00:19:49,829 What we have is then a whole

    679 00:19:49,839 –> 00:19:51,400 set of integral

    680 00:19:51,410 –> 00:19:53,280 values where the only condition

    681 00:19:53,290 –> 00:19:55,180 is that our step function

    682 00:19:55,189 –> 00:19:57,089 is always below

    683 00:19:57,189 –> 00:19:58,810 the real function F.

    684 00:20:01,239 –> 00:20:02,900 The general idea is now

    685 00:20:02,910 –> 00:20:03,599 OK.

    686 00:20:03,609 –> 00:20:05,500 So this integral value for

    687 00:20:05,510 –> 00:20:07,459 the step function is always

    688 00:20:07,469 –> 00:20:08,939 smaller than the

    689 00:20:08,949 –> 00:20:10,859 actual integral value for

    690 00:20:10,869 –> 00:20:12,150 the function F.

    691 00:20:12,160 –> 00:20:13,609 And we should get closer

    692 00:20:13,619 –> 00:20:15,160 and closer to this value

    693 00:20:15,170 –> 00:20:17,040 if we choose a finer and

    694 00:20:17,050 –> 00:20:18,890 finer decomposition of the

    695 00:20:18,900 –> 00:20:20,800 Y axis, so

    696 00:20:20,810 –> 00:20:22,359 we approximate something

    697 00:20:22,369 –> 00:20:23,670 with the set here.

    698 00:20:23,680 –> 00:20:25,040 And of course, this should

    699 00:20:25,050 –> 00:20:26,859 be the supremum of the

    700 00:20:26,869 –> 00:20:27,319 set.

    701 00:20:28,300 –> 00:20:29,579 In fact, this is what we

    702 00:20:29,589 –> 00:20:31,069 choose as the

    703 00:20:31,079 –> 00:20:32,430 definition of the

    704 00:20:32,439 –> 00:20:34,209 integral for our function

    705 00:20:34,219 –> 00:20:34,800 F.

    706 00:20:34,810 –> 00:20:36,560 So we have F d mu

    707 00:20:37,099 –> 00:20:39,069 and also this is our measure

    708 00:20:39,079 –> 00:20:39,900 space X.

    709 00:20:40,750 –> 00:20:42,540 So the integral of a nonne

    710 00:20:43,199 –> 00:20:45,089 measurable map

    711 00:20:45,099 –> 00:20:47,089 F is given

    712 00:20:47,099 –> 00:20:48,859 by the supremum of

    713 00:20:48,869 –> 00:20:50,589 all integral values for

    714 00:20:50,599 –> 00:20:52,130 step functions that

    715 00:20:52,140 –> 00:20:53,989 lie below the function

    716 00:20:54,000 –> 00:20:54,479 F.

    717 00:20:55,390 –> 00:20:56,609 And that is now the

    718 00:20:56,619 –> 00:20:58,119 definition of the Lebesgue

    719 00:20:58,439 –> 00:20:59,089 integral.

    720 00:20:59,770 –> 00:21:01,579 And we also see this is well

    721 00:21:01,589 –> 00:21:02,910 defined the

    722 00:21:02,920 –> 00:21:04,650 supremum of a

    723 00:21:04,660 –> 00:21:06,550 set in the real numbers always

    724 00:21:06,560 –> 00:21:07,390 exists.

    725 00:21:07,619 –> 00:21:09,060 In the worst case, it would

    726 00:21:09,069 –> 00:21:10,030 be infinity.

    727 00:21:11,010 –> 00:21:12,400 Therefore, in addition, we

    728 00:21:12,410 –> 00:21:14,300 also have another definition

    729 00:21:14,800 –> 00:21:16,599 F is called mu-integrable

    730 00:21:19,239 –> 00:21:20,780 if the integral is

    731 00:21:20,790 –> 00:21:21,619 finite.

    732 00:21:22,920 –> 00:21:24,739 So it’s not the symbol

    733 00:21:24,750 –> 00:21:25,699 infinity.

    734 00:21:27,770 –> 00:21:29,219 And there we have it, this

    735 00:21:29,229 –> 00:21:30,900 is our result for today,

    736 00:21:30,910 –> 00:21:32,530 the Lebesgue integral in

    737 00:21:32,540 –> 00:21:34,219 complete generality

    738 00:21:34,280 –> 00:21:35,750 because you see the only

    739 00:21:35,760 –> 00:21:37,420 thing we needed here

    740 00:21:37,550 –> 00:21:38,969 was a measure space

    741 00:21:38,979 –> 00:21:39,599 X.

    742 00:21:40,010 –> 00:21:41,430 So we only needed to know

    743 00:21:41,439 –> 00:21:43,030 how to measure the

    744 00:21:43,040 –> 00:21:45,030 volumes of these sets

    745 00:21:45,040 –> 00:21:46,489 on the X axis.

    746 00:21:47,160 –> 00:21:48,689 However, what we use is that

    747 00:21:48,699 –> 00:21:50,339 we map into the real

    748 00:21:50,349 –> 00:21:51,069 numbers.

    749 00:21:51,079 –> 00:21:52,910 So indeed, here we have R

    750 00:21:52,920 –> 00:21:54,250 on the Y axis.

    751 00:21:54,300 –> 00:21:55,770 Therefore, we can do this

    752 00:21:55,780 –> 00:21:57,380 decomposition, we can’t do

    753 00:21:57,390 –> 00:21:58,310 this decomposition on the

    754 00:21:58,319 –> 00:22:00,150 X axis because there’s no

    755 00:22:00,160 –> 00:22:02,089 order is nothing more than

    756 00:22:02,099 –> 00:22:03,630 just the measure.

    757 00:22:03,880 –> 00:22:05,469 But on the y axis, we can

    758 00:22:05,479 –> 00:22:05,869 do it.

    759 00:22:06,949 –> 00:22:08,119 Now, we should always keep

    760 00:22:08,130 –> 00:22:09,920 in mind the Lebesgue integral

    761 00:22:09,930 –> 00:22:11,849 is just defined within

    762 00:22:11,859 –> 00:22:13,689 supremum where we use

    763 00:22:13,699 –> 00:22:15,369 simple functions to

    764 00:22:15,380 –> 00:22:17,010 define the integral first.

    765 00:22:18,079 –> 00:22:19,699 So this was a long video

    766 00:22:19,709 –> 00:22:21,569 today and I hope you

    767 00:22:21,579 –> 00:22:22,699 learned something here.

    768 00:22:23,530 –> 00:22:25,270 Of course, we had to define

    769 00:22:25,280 –> 00:22:26,290 a lot of stuff.

    770 00:22:26,380 –> 00:22:28,270 But in the end, you saw

    771 00:22:28,599 –> 00:22:30,400 it is just a simple function

    772 00:22:30,410 –> 00:22:31,670 where we can write down the

    773 00:22:31,680 –> 00:22:33,020 integral immediately.

    774 00:22:33,030 –> 00:22:34,540 And then the

    775 00:22:34,550 –> 00:22:36,469 general integral is just

    776 00:22:36,479 –> 00:22:38,150 an approximation concept.

    777 00:22:39,010 –> 00:22:40,540 In the next video, I will

    778 00:22:40,550 –> 00:22:41,790 continue with the Lebesgue

    779 00:22:41,800 –> 00:22:43,709 integral and also

    780 00:22:43,719 –> 00:22:45,079 write down a lot of

    781 00:22:45,089 –> 00:22:46,939 properties we have for it.

    782 00:22:47,780 –> 00:22:49,589 And afterwards, you will

    783 00:22:49,599 –> 00:22:51,489 finally see why the Lebesgue

    784 00:22:51,609 –> 00:22:53,560 integral is so powerful,

    785 00:22:54,619 –> 00:22:56,180 then have a nice day

    786 00:22:56,250 –> 00:22:57,890 and see you next time.

    787 00:22:58,310 –> 00:22:58,920 Bye.

  • Quiz Content

    Q1: Let $(X, \mathcal{A}, \mu)$ be a measure space and $\chi_A: X \rightarrow \mathbb{R}$ the indicator function for $A \in \mathcal{A}$. What is the integral of $\chi_A$ w.r.t. $\mu$ (by definition)?

    A1: $I(\chi_A) = \mu(A^c)$

    A2: $I(\chi_A) = 1-\mu(A)$

    A3: $I(\chi_A) = \mu(A)$

    A4: $I(\chi_A) = \mu(A^c) + \mu(A)$

    Q2: Let $(X, \mathcal{A}, \mu)$ be a measure space. Which property does a simple function $f: X \rightarrow \mathbb{R}$ not fulfil?

    A1: ${ f(x) \mid x \in X }$ is finite

    A2: $f$ is measurable

    A3: $f(x) \in \mathbb{R}$ for all $x \in X$

    A4: $f^{-1}({\mathbb{R}}) = X$

    A5: ${ f(x) \mid x \in X } = \mathbb{R}$

    Q3: Consider $(X, \mathcal{A}) = (\mathbb{R}, \mathcal{B}(\mathbb{R}) )$. Is the following function $f: \mathbb{R} \rightarrow \mathbb{R}$ a simple function? $$ f(x) := \sum_{i=1}^{100} i \cdot \chi_{[0,i]}(x) $$

    A1: Yes.

    A2: No.

    Q4: Consider $(X, \mathcal{A}) = (\mathbb{R}, \mathcal{B}(\mathbb{R}) )$. Is the following function $f: \mathbb{R} \rightarrow \mathbb{R}$ a simple function? $$ f(x) := \chi_{[0,1]}(x) - \chi_{\mathbb{R}}(x) $$

    A1: Yes, it is.

    A2: No, it isn’t.

    Q5: Let $(X, \mathcal{A}, \mu)$ be a measure space. Which of the following statements is not a property of the Lebesgue integral (w.r.t. $\mu$) of non-negative simple functions as a map $I: \mathcal{S}^+ \rightarrow [0,\infty]$.

    A1: $I(f+g) = I(f) + I(g)$.

    A2: $I(\alpha f) = \alpha I(f)$ for $\alpha \geq 0$.

    A3: $f < g ~~\Rightarrow ~~ I(f) < I(g)$

    Q6: Let $(X, \mathcal{A}, \mu)$ be a measure space. Can we define the Lebesgue integral for every measurable function $f: X \rightarrow [0,\infty)$?

    A1: Yes and we have $\int_X f , d \mu \in \mathbb{R}$.

    A2: Yes and we have $\int_X f , d \mu \in \mathbb{C}$.

    A3: Yes and we have $\int_X f , d \mu \in [0,\infty]$.

    A4: No, there are exceptions.

  • Last update: 2024-10

  • Back to overview page