CONTENTS PREFACE... PROLOGUE: WHAT 15 FUNCTION? Exercise . 1 LET'S DIFFERENTIATE A FUNCTION! Approximating with Funetfons Calculating the Relative Esror. The Derivative in Action! Step 1 Step 2. Step 3, : Caleulating the Derivative .- Calculating the Derivative of a Constant, Linear, or Quadratic Function . Summary Bxercises . 2 LET'S LEARN DIFFERENTIATION TECHNIQUES! ‘The Sum Rule of Differentiation . ‘The Product Rule of Differentiation Differentiating Polynomials. Finding Maxima and Minima Using the Mean Value Theorem. . Using the Quotient Rule of Differentiation Calculating Derivatives of Composite Functions . Calculating Derivatives of Inverse Functions Exereises .... 3 LET'S INTEGRATE A FUNCTION! lusteating the Fundamental Theorem of Calculus ‘Step 1—When the Density Is Constant. ‘Step 2~-When the Density Changes Stepwise ‘Step 3—When the Density Changes Continuously . Step 4—Review of the Imitating Linear Function. Step 5—Approsimation + Exact Value Step 6—p(x) Is the Derivative of q(x) “ 18 16 27 32 39 “7 82 83 84 55, 88 2-89 +90Using the Fundamental Theorem of Calculus. fanaeits <3 Summary... ECsainn 33 o bdeSo 059 | A Strict Explanation of Sep 3. Lea | Using Integral Formulas i fee ccnn aa ccmeeese) Applying the Fundamental Theorem : 101 Baynly Caretews css tcecamsviveserson sve eis comm easyatweess AO Demand Curve 103 Review of the Fundamental Theorem of Caleulus. . eetry Formula of the Substitution Rule of Integration. aun The Power Rae of tegration ae eee) BROPCISCS 6... eeeeeeeeeetesesevesvereenertese oman 4 | LET'S LEARN INTEGRATION TECHNIGUES!. oe sees | Using Trigonometric Functions... fee eee Ln Using Integrals with Trigonometric Functions 15 Using Exponential and Logarithmic Functions... aloes iibt Generalizing Exponential and Logarithmic Functions 21-135 Summary of Exponential and Logarithmic Functions +140 | More Applications of the Fundamental Theorem Taga | Integration by Parts, sas Exercises on mer) | 5 | LET'S LEARN ABOUT TAYLOR EXPANSIONS! 145 Imitating with Polynomials .....0.060c0cesceeeeeeeseeees 147 How to Obtain a Taylor Expansion 158 ‘Taylor Expansion of Various Functions - 2160 What Does Taylor Expansion Tell Us?. . se daewussar a6 Exercises, oe scenes cea 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION... -179 What Are Multivariable Funetions?. -180 The Bastes of Variable Linear Functions 184 Partial Differentiation ...... : is Definition of Partial Differentiation +196 ‘Total Differentials 1197 Conditions for Extrema... 1199 ‘Applying Partial Differentiation to Economies. 20a The Chain Rule .. 2206 Derivatives of implicit Functions 2218 Baercises - 2218 ul contentsEPILOGUE: WHAT IS MATHEMATICS FOR?. . A | SOLUTIONS TO EXERCISES .... non | Protogue. Chapter L.......-. Chapter 2. . Chapter 3... Chapter 4 Chapter 5 Chapter 6... 8 MAIN FORMULAS, THEOREMS, AND FUNCTIONS | COVERED IN THIS BOOK Linear Equations (Linear Functions) | Disferentiation . . Derivatives of Popular Functions Integrals. . : ‘Taylor Expansion Partial Derivatives | INDex... 219) 225 225 1225 225 2-226 227 228 =-228 =-281 aan 2-231 +082 2-234 234 235PREFACE ‘There are some things that only manga can do. ‘You have just picked up and opened this book. You must be one of the following types of people. The first type is someone who just loves manga and thinks, “Calculus illustrated with manga? Awesome!” If you are this type of person, you should immediately take this book to the cashier— you won't regret it. This is a very enjoyable manga title. It's no ‘surprise—Shin Togami, a popular manga artist, drew the manga, and Becom Ltd., a real manga production company, wrote the “But, manga that teaches about math has never been very enjoyable,” you may argue. That's true. In fact, when an editor at Ohmsha asked me to write this book, I nearly turned down the opportunity. Many of the so-called “manga for education” books are quite disappointing. They may have lots of illustrations and large pictures, but they aren't really manga. But after seeing a sample from Ohmsha (it was The Manga Guide to Statistics), 1 totally changed my mind. Unlike many such manga guides, the sample was enjoyable enough to actually read. The editor told me that my book would be like this, too—so I accepted his offer. In fact, I have often thought that I might be able to teach mathemat- ics better by using manga, so I saw this as a good opportunity to put the idea into practice. I guarantee you that the bigger manga freak you are, the more you will enjoy this book. So, what are you waiting for? Take it up to the cashier and buy it already! ‘Now, the second type of person is someone who picked up this book thinking, “Although I am terrible at and/or allergic to calcu- lus, manga may help me understand it.” If you are this type of per- son, then this is also the book for you. It is equipped with various rehabilitation methods for those who have been hurt by calculus in the past. Not only does it explain calculus using manga, but the way it explains calculus is fundamentally different from the method used in conventional textbooks. First, the book repeatedlyXi PREFACE: presents the notion of what calculus really does. You will never understand this through the teaching methods that stick to limits (or c-8 logic). Unless you have a clear image of what calculus really does and why it is useful in the world, you will never really under- stand or use it freely. You will simply fall into a miserable state of memorizing formulas and rules. This book explains all the formu- las based on the concept of the first-order approximation, helping you to visualize the meaning of formulas and understand them easily. Because of this unique teaching method, you can quickly and easily proceed from differentiation to integration. Further- more, I have adopted an original method, which is not described in ordinary textbooks, of explaining the differentiation and integra- tion of trigonometric and exponential functions—usually, this is all Greek to many people even after repeated explanations. This book also goes further in depth than existing manga books on calculus do, explaining even Taylor expansions and partial dif- ferentiation. Finally, I have invited three regular customers of calculus—physics, statistics, and economics—to be part of this book and presented many examples to show that calculus is truly Practical. With all of these devices, you will come to view caleulis not as a hardship, but as a useful tool. Iwould like to emphasize again: All of this has been made possible because of manga. Why can you gain more information by reading a manga book than by reading a novel? It is because manga is visual data presented as animation. Calculus is a branch of mathematics that describes dynamic phenomena—thus, calew- lus is a perfect concept to teach with manga. Now, turn the pages and enjoy a beautiful integration of manga and mathematics. HIROYUKI KOJIMA NovemBe 2005 Nowe: For ease of understanding, some figures are not drawn to scale,PROLOGUE: WHAT IS A FUNCTION?i Wi, UST THINK—ME, NORIKO HIKIMA, A JOURNALIST! MY R HERE! 37'S A SMALL NEWSPAPER AND “JUST A BRANCH ‘STILL A JOURNALIST!A NEWSPAPER DISTRIBUTOR? YOU'RE LOOKING | Af 5 veer || OUEE ROR, Sean orrce? evenvbooy nsraces WHAT 15 A FUNCTION? 3HERE GOES NOTHING! GOOD MORNING!WILL YOU LEAVE Tr, PLEASE? OH, YOU HAVE BEEN ASSIGNED HERE TODAY. THE BIG Guy THERE 1S FUTOSHI MASUL, My ONLY SOLDIER.YES! THINKING: ABOUT FACTS. AFACT IS SOMEHOW RELATED TO ANOTHER FACT. UNLESS YOU UNDERSTAND THESE RELATIONSHIPS, YOU WON'T BE A REAL WHAT 56 A FUNCTION? 7YES! THATS TRUE-I'VE STUDIED LITERATURE SINCE TWAS A JUNIOR IN HIGH SCHOOL. CATCHING UP TO Do, THEN. LET'S BEGIN WITH FUNCTIONS. A FUNCTION DESCRIBES A RELATION, CAUSALITY, OR CHANGE. A FUNCTION IS A CORRELATION. YOU CAN THINK OF THE WORLD ITSELF AS ‘ONE BIG FUNCTION. _/) AS JOURNALISTS,y= SKI? FOR EXAMPLE, DID_YOU KNOW A ASSUME x FUNCTION IS OFTEN AND y ARE EXPRESSED AS ANIMALS. Animal x — [F]— [Animary ASSUME x IS A FROG, IF YOU PUT THE FROG INTO BOX f AND CONVERT IT, TADPOLE y COMES OUT BUT, UH. WHAT 16, f2 THE f STANDS FOR FUNCTION, NATURALLY. J unction S15 USED TO SHOW THAT THE VARIABLE y HAS A PARTICULAR RELATIONSHIP TO x. AND WE CAN ‘ACTUALLY USE ANY] LETTER INSTEAD OF Sf. Waar 15 A FINCTON? @IN THIS CASE, f EXPRESSES THE RELATIONSHIP ‘OR RULE BETWEEN *A PARENT" AND "AN OFFSPRING." AND THIS, RELATIONSHIP IS TRUE OF ALMOST ANY ANIMAL. IF x IG ABIRD, y IS A CHICK. ‘OKAY! NOW LOOK AT THIS. THE RELATIONSHIP. BETWEEN INCOMES }__ AND EXPENDITURES | CAN BE SEEN ASA 1 row wHen THE SALES ATA COMPANY GO UP, ‘THE EMPLOYEES 10 FeoLosie MOUNTAINS GOES DOWN BY ABOUT 05°C EACH TIME YOU ‘GO UP 100 METERS, DOESNT IT?We Have PLexry DO YOU GET IT? WE TRIN ABOUT THESE ARE SURROUNDED BY THINGS QUIETLY. PINCTONS IT'S A SMALL OFFICE, BUT I HOPE YOU WILL DO YOUR WHAT 6 A FUNCTION? 1ee ‘OH, LUNCH 1S HERE: ALREADY? WHERE IS MY BEEF BOWL? ]FUTOSHI, LUNCH NOT YET? PLEASE HASN'T COME WAKE ME UP WHEN LUNCH IS HERE. 222. HAS LUNCH NO, FUTOSHI, WE HAVE A COMPF. NEW. We] NO, NOT Yer. i 2 PROLOGUETABLE |: CHARACTERISTICS OF FUNCTIONS Buber ALEILATION Rar Causality | The frequency of cricket's chirpis | When we graph these determined by temperature. We can functions, the result is express the relationship between a straight line. That's xy chirps per minute of a cricket at hy we call them linear temperature x°C approximately as functions. y=g(x)=7x-90 y. tot x=27° 7% 27-30 ‘The result is 159 chirps a minute Changes | The speed of sound y in meters per sec- cond (mis) in the air at x°C is expressed as (x) =0.6x +331 ALISC, y=0(15)=0.615 +331 = 340 mis At-8°, ° x y=v(-8)=0.6x(-5)+331=328 m/s Unit ‘Converting x degrees Fahrenheit 'F) into Conversion | y degrees Celsius (°C) =2¢<-20) 9 So now we know 50°F is equivalent to | = S50 5 : 5(50~s2)-10°C al ) ‘Computers store numbers using a binary system (Is and 0s). A binary number with x bits (or binary digits) has the potential to store y numbers. y=b(x)=2" (This is described in more detail on page 131.) ‘The graph is an expo- nential fun: WHAT 36 A FUNCTION? 13THE GRAPHS OF SOME FUNCTIONS CANNOT BE EXPRESSED BY STRAIGHT LINES OR CURVES WITH A REGULAR SHAPE. ‘The stock price P of company A in month x in 2009 is, y= PIX) Yen 1 2 3 4 5 6 Month P(x) cannot be expressed by a known function, but it is still a function. Ifyou could find a way to predict P{7), the stock price in July, you could make a big profit. COMBINING TWO OR MORE FUNCTIONS IS CALLED ‘THE COMPOSITION OF FUNCTIONS.” COMBINING FUNCTIONS ALLOWS US TO EXPAND THE RANGE OF CAUSALITY. ‘Acomposite function offandg [9]—rovey) EXERCISE 1. Find an equation that expresses the frequency of z chirps/minute of a cricket at x°F. 14 PROLOGUE1 LET'S DIFFERENTIATE A FUNCTION! PTV TrO Teeter yyAPPROXIMATING WITH FUNCTIONS ALL RIGHT, 7M DONE FOR THE NORIKO, 1 HEARD BUT..YOU'RE A POSH TTALIAN FINISHED RESTAURANT JUST ALREADY? OPENED NEARBY. I'S NOT EVEN WOULD YOU LIKE NOON. TTALIAN FOOD. LETS Go! 16 CHAPTER 1 LET'S DIFFERENTIATE A FUNCTION!| To: eonons ‘SUBJECT: TODAY'S HEADLINES A BEAR RAMPAGES IN A HOUSE AGAIN THE REPUTATION IMPROVES IN THE PRE? OF SANDA-CHO WATERMELONS FECTURE NO INJURIES aN AH. THAT'S \ imposeiete. APPROKIMATING WITH FUNCTIONS 17I KNEW IT. I DON’T WANNA WORK HERE! BUT I WILL TRAIN YOU WELL SO THAT YOU CAN BE ACCEPTED AT THE MAIN OFFICE. 18 CHAPTER | LET'S DIFFERENTIATE A FUNCTION!BY THE WAY, DO YOU THINK THE JAPANESE ECONOMY 15 STILL EXPERIENCING DEFLATION? THE GOVERNMENT REPEATEDLY SAID THAT THE ECONOMY WOULD RECOVER. MUST FIRST ASK FH A TRUE JOURNALIST HIMSELF, WHAT DO INK $0. 1 FEEL My DAILY LIFE. i BUT IT TOOK A LONG TIME UNTIL SIGNS OF RECOVERY APPEARED. HAVE A BAD FEELING ABOUT THIS... APPROKMMATING WITH FUNCTIONS 1@HERE We IF YOU CAN APPROXIMATE WHAT USE A LINEAR YOU WANT TO KNOW WITH A EXPRESSION: | | NOW, WHAT WE WANT SIMPLE FUNCTION, YOU CAN SEE TO KNOW MOST IS IF THE ANSWER MORE CLEARLY. PRICES ARE GOING 60 IF ais NEGATIVE, WE KNOW THAT DEFLATION 15. 3008 E STILL CONTINUING. (Year) jo ‘Turned to inflation Still in deflation 20 CHAPTER! LET'S DIFFERENTIATE A FUNCTION!THAT'S RIGHT. YOU ARE A QUICK STUDY, NOW, LET'S DO THE REST AT THE TTALIAN RESTAURANT. FUTOSHI, WE'RE LEAVING FOR LUNCH. DONT EAT TOO MANY ‘SNACKS. SPEAKING OF SNACKS, DO YOU KNOW ABOUT JOHNNY FANTASTIC, THE ROCKSTAR WHOSE BOOK ON DIETING HAS BECOME A BEST SELLER? APPROXIMSTING WITH FUNCTIONS 21BUT HE SUDDENLY BEGAN TO GAIN WEIGHT AGAIN by ALTHOUGH HIS. GAIN HAS: AGENT WARNED HIM ABOUT IT, ASSED I WHETHER JOHNNY'S WEIGHT GAIN IS REALLY SLOWING DOWN LIKE HE SAID. HE WAS CERTAIN. NOW WHAT HIS AGENT WANTS TO YOU'RE RIGHT. NOW, LET'S IMITATE HIS WEIGHT, GAIN WITH yrax’+bx+e ax’ tox +e Weight (kg) Weight (kg) 22 CHAPTER! LETS DIFFERENTIATE A FUNCTION!WEIGHT GAIN IS ACCELERATING. SLOWING DOWN. WEIGHT GAIN IS AND IF LET'S ASSUME YOU WANT TO KNOW. HOW TIGHT EACH CURVE 15. WE CAN APPROXIMATE EACH CURVE WITH A CIRCLE. THERE ARE LOTS OF TIGHT CURVES: AROUND HERE. IF a IS POSITIVE, HIS WEIGHT GAIN IS ACCELERATING. SLOWING DOWN. IS NEGATNE, IT'S Goop! YOU'RE DOING welt. Tt EH, I DONT. REALLY CARE ‘ABOUT THAT. APPROXIMATING WITH FUNCTIONS 23LOOK. ASSUME THE CURVATURE OF THE ROAD IS ON THE CIRCUMFERENCE OF A CIRCLE WITH RADIUS R. THE SMALLER RIG, THE y=\R'-(x-a)' +b (x-a) +(y-by =k LET'S IMITATE IT WITH THE FORMULA FOR A CIRCLE WITH RADIUS R CENTERED AT POINT (a, b). i Ce 24 CHAPTER 1 LET'S DIFFERENTIATE A FUNCTION!WELL, THAT'S THE TTALIAN RESTAURANT WE WANT TO GO TO. IT'S STILL $0 FAR AWAY. LET'S DENOTE THIS ACCIDENT SITE WITH POINT P. AND LET'S THINK OF THE ROAD AS A GRAPH OF THE, FUNCTION fix) = x7. ACCIDENT APPROXIMATING WITH FUNCTIONS 25Italian restaurant THE LINEAR FUNCTION THAT AT POINT P APPROXIMATES THE FUNCTION THE SLOPE RISES Six) = x° COUR ROAD) AT x = 215 4 KILOMETERS VERTICALLY ‘glx) = 4x - 4. THIS EXPRESSION FOR EVERY | KILOMETER CAN BE USED TO FIND OUT, TT GOES HORIZONTALLY. IN FOR EXAMPLE, THE SLOPE AT REALITY, MOST OF THIS ROAD THIS PARTICULAR POINT. 1S NOT 50 STEEP. ‘THE REASON IS GNEN ON PAGE 30. FUTOSHIP WE'VE HAD AN ACCIDENT. WILL YOU HELP use | THE ACCIDENT Sie? I'S POINT P. 26 CHAPTER LET'S DIFFERENTIATE A FUNCTION! WHAT FUNCTION SHOULD 1 USE TO APPROXIMATE THE INSIDE OF YOUR HEAD?CALCULATING THE RELATIVE ERROR THE RELATIVE ERROR GIVES THE RATIO OF THE v DIFFERENCE BETWEEN THE WHILE we wart For i i VALUES OF f(x) AND g(x) TO FUTOSHL, PLL TELL THE VARIATION OF x WHEN x YOU ABOUT RELATIVE J 1) 1S CHANGED. THAT IS... ERROR, WHICH IS ‘ ALSO IMPORTANT. Our Our original approximating function function 1 t Difference between flx) and g(x) Relative error = ee ‘Change of x CALCULATING THE RELATWE ERROR 27ASSUME THAT x EQUALS 2 AT THE POINT WHERE WE ARE NOW AND THAT THE: DISTANCE FROM HERE TO THE RAMEN SHOP 15 O18. NOW, ASSUME THE POINT WHERE T AM STANDING IS (0.01 FROM P. 90 THE DIFFERENCE IS f(2.1) ~ (2.1) = 0.01, AND THE RELATIVE ERROR IS 0.01 /0.1 = 0.1 GO PERCENT). 28 CHAPTER LET'S DIFFERENTIATE A FUNCTION!CHANGE x BY 0.01: x = 2 BECOMES x= 2.01. sssor flan)-g(200- 40401404 0.001 RELATIVE ERROR 2 pool = 0.0! (1%) As the variation approaches 0, the relative error also approaches 0. Variation of fix) oe) Error Relative xfrom 2 error 1 9 8 2 100.0% o1 4.41 44 0.01 10.0% 0.01 4.0401 4.04 0.0001 1.0% 0.001 4.004001 4.004 0.000001 0.1% + + 0 ° IN OTHER WORDS, THE CLOSER I STAND TO THE ACCIDENT SITE, THE BETTER g(x) IMTATES fix). THE RELATIVE ERROR FOR THIS POINT IS SMALLER THAN FOR THE RAMEN SHOP. CALCULATING THE RELATIVE ERROR 20THAT'S NOT $0 SIRORENG, Be SREENSY WibERSTAND a. DERIVATIVES. $0, THE restaurant || BE STRAIGHT WITH YES. TODAY WE WILL HAVING THE || ME! WE'RE GONNA EAT AT THE RAMEN ‘SMALLEST EAT AT THE RAMEN ‘SHOP, WHICH 1S. ‘SHOP, ARENT WE? CLOSER TO POINT P. ‘THE APPROXIMATE LINEAR FUNCTION IS SUCH THAT ITS RELATIVE ERROR WITH RESPECT TO THE ORIGINAL FUNCTION IS LOCALLY ZERO. 20, AS LONG AS LOCAL PROPERTIES ARE CONCERNED, WE CAN DERNE THE CORRECT RESULT BY USING THE APPROXIMATE LINEAR FUNCTION FOR THE ORIGINAL FUNCTION. ‘986 PAGE 39 FOR THE DETAILED CALCULATION. 30 CHAPTER LETS DIFFERENTIATE A FUNCTION!NORIKO, WE CAN ALSO ESTIMATE THE COST- EFFECTIVENESS OF TV COMMERCIALS: USING APPROXIMATE FUNCTIONS. SIGH. I LIKE RAMEN, || BUT I WANTED To EAT ITALIAN FOOD. CALCULATING THE RELATING ERROR 31THE DERIVATIVE IN ACTION! LET'S CONSIDER WHETHER ONE OF THEIR EXECUTIVES INCREASED OR DECREASED THE AIRTIME (OF THE COMPANY'S TV COMMERCIAL TO RAISE YOU KNOW THE BEVERAGE \ MANUFACTURER THE PROFIT FROM ITS AMALGAMATED POPULAR PRODUCTS. cOLAe TLL DO mt WILL OKAY, I GUESS, WHEN I WORKED AT THE MAIN OFFICE, ONLY WORK HARD. ONE MAN SOLVED THIS PLEASE TELL ME PROBLEM. HE IS NOW A THE STORY. HIGH-POWERED... ‘ASSUME AMALGAMATED COLA AIRS ITS TV COMMERCIAL x HOURS PER MONTH, TT IS KNOWN THAT THE PROFIT FROM INCREASED SALES DUE TO x HOURS OF COMMERCIALS 1S: Six) =20Vx (IN HUNDREDS OF MILLION YEN). 32 CHAPTER LET'S DIFFERENTIATE A FUNCTION!AMALGAMATED COLA NOW AIRS THE TV COMMERCIAL FOR 4 HOURS PER MONTH. AND SINCE S(4)=204/4 = 40, THE COMPANY MAKES A'PROFIT OF 4 BILLION YEN. TPMINUTE COMMERCIAL = THE FEE FoR THE 1V COMMERCIAL IS Sel 10 MILLION YEN PER MINUTE. ‘ls} =20V8 HUNDRED MILLION YEN I-MIN COMMERCIAL = ¥10 MILLION NOW, A NEWLY APPOINTED EXECUTIVE HAS DECIDED TO RECONSIDER THE AIRTIME OF THE TV COMMERCIAL. DO YOU THINK HE WILL INCREASE THE AIRTIME OR DECREASE IT? be Je [BN 4) ‘THE DERWATNE IN ACTION! 33SINCE IT'S IMPOSSIBLE JO IMITATE THE WHOLE FUNCTION WITH A LINEAR FUNCTION, WE WILL IMITATE IT IN THE VICINITY (OF THE CURRENT AIRTIME OF x=4, fa)-208k HUNDRED MILLION YEN SINCE f(x) = 200% HUNDRED MILLION YEN 1S. A COMPLICATED FUNCTION, LET'S MAKE A SIMILAR LINEAR FUNCTION TO ROUGHLY ESTIMATE THE RESULT. WE WILL DRAW A, TANGENT LINE" TO. \ THE GRAPH OF So) =20N% L ar oe * Here is the calculation of the tangent line. (See also the explanation of the derivative on page 39.) For f(x) = 20Vx,S(4) is given as follows. S(4+6)-S(4) _ 20/46 -~20%2 _ 94 (V4 +6 -2)x(V4 = ex(lase+3) When ¢ approaches 0, the denominator of © J4+e+2— 4. Therefore, © ~ 20 Thus, the approximate linear function g(x) =5(x—4)+40=5x +20 34 CHAPTER 1 LET'S DIFFERENTIATE A FUNCTION!LARGE—FOR EXAMPLE, CANNOT BE USED. IN AIRTIME OF THE TV. BE A SMALL AMOUNT, DECREASE. IF THE CHANGE IN x IS. HOUR—THEN g(x) DIFFERS FROM Jix) TOO MUCH AND IN REALITY, THE CHANGE COMMERCIAL MUST ONLY EITHER AN INCREASE OR A aN IF YOU CONSIDER AN INCREASE OR DECREASE (OF; FOR EXAMPLE, 6 MINUTES (1 HOUR), THIS APPROXIMATION CAN BE USED, BECAUSE THE RELATIVE ERROR 1S SMALL WHEN THE CHANGE IN x IS SMALL. We FIND THAT IN THE VICINITY OF x= AN INCREASE OF HOURS, fix) CAN BE SAFELY || 6 MINUTES BRINGS: APPROXIMATED AS ROUGHLY || A PROFIT INCREASE glx) = 5x +20, OF ABOUT 5 x 04 = (0.5 HUNDRED MILLION YEN. THE FACT THAT THE COEFFICIENT OF x IN glx) |S 5 MEANS A PROFIT INCREASE (OF 5 HUNDRED MILLION YEN PER HOUR. $0 IF THE CHANGE 15 ONLY 6 MINUTES (01 HOUR), THEN WHAT HAPPENS? “THAT'S RIGHT. BUT, HOW MUCH DOES Ir COST TO INCREASE THE AIRTIME OF THE COMMERCIAL BY 6 MINUTES? THE FEE FOR THE INCREASE |S 6 x O1 = 0.6 HUNDRED MILLION YEN. TF INSTEAD, THE AIRTIME: IS DECREASED BY 6 MINUTES, THE PROFIT DECREASES ABOUT 05 BILLION YEN. BUT SINCE YOU DON'T HAVE TO PAY THE FEE OF (0.6 HUNDRED MILLION YEN, THE DERNATNE IN ACTON! 35THE ANSWER IS...THE COMPANY, DECIDED TO DECREASE THE COMMERCIAL TIME! PEOPLE USE FUNCTIONS LEMS. BY THE WAY, WHO 15 THE TO SOLVE PROBI IN BUSINESS AND LIFE IN /(~ MAN THAT SOLVED THIS; THE REAL WORLD. PROBLEM? THAT'S TRUE. WHETHER THEY ARE CONSCIOUS OF FUNCTIONS OR NOT. 36 CHAPTER LET'S DIFFERENTIATE A FUNCTION!AS I EXPECTED..SOLVING MATH PROBLEMS HAS NOTHING TO DO WITH BEING A HIGH-POWERED JOURNALIST. SAID HE WAS HIGH-POWERED, DIDN'T you? YOU'RE KIDDING! THE DERNATE INACTION! 37THis 16 ABSURD! IWONT GNE UP [1] LUNCHTIME 1S OVER. [|] LETS FIK THE CARY Do WITH BEING A JOURNALIST...CALCULATING THE DERIVATIVE Let's find the imitating linear function g(x) = kx + Lof function fix) at x = a. We need to find slope k. © a(x k(x-a)+JS(a) (g(x) coincides with fla) when x = a.) Now, let's calculate the relative error when x changes from x = a to Difference between f and g after x has changed ‘Change of x from S(a+e)-g(a+e) i g(a+e)=k(a+e-a)+ F(a) S(are)-(ke+ s(a)) =ke+ f(a) ‘ ‘When & approaches 0, Slare)-$la) a the relative error also approache: pp Ll@+e)-Fi SSPE Os ara N po when 2+ 0. (The lim notation expresses the operation that obtains the value when ¢ approaches 0.) Linear function ®, or g(x), with this k, is an approximate fuinction of fix). kis called the differential coefficient of fix) at x= a. tim £(4*2)- F(a) Slope of the line tangent to y =/(x) at 5 € any point (a, f(a). We make symbol f’ by attaching a prime to f. S(a+e)-S(a) f(a) is the slope of the line tangent to Fie) ae y=fixj atx= Letter a can be replaced with x. Since J’ can been seen as a function of x, it is called “the function derived from function J,” or the derivative of function f. CALCULATING THE DERVATIVE. 32CALCULATING THE DERIVATIVE OF A CONSTANT, LINEAR, OR QUADRATIC FUNCTION 1. Let's find the derivative of constant function fix) = a. The differential coefficient of fix) at x = ais tim S12 # 9)" F(@) | ie 22 -ttmo=0 Thus, the derivative of fix) is f(x) = 0. This makes sense, since our function is constant—the rate of change is 0. Nowe The differential coefficient of flx) at x = ais often simply called the derivative of fix} at x = a, or just f(a). 2. Let's calculate the derivative of linear function f[x) = ax +P. The deriva- tive of fix) at x= ais lim. Ji (a+e)-s(a) _,,2(a46)+A-(aa+ p) Thus, the derivative of fix) is fx) = @, a constant value. This result should also be intuitive—linear functions have a constant rate of change by definition. 3 Let's find the derivative of fix) = x", which appeared in the story. The dif- ferential coefficient of fix) at x= ais my F(A) $4) _ ay (at 6) lim(2a+6)=2a ‘Thus, the differential coefficient of fix) at x = a is 2a, or fa) = 2a. Therefore, the derivative of fix) is f(x) = 2x. SUMMARY The calculation of a limit that appears in calculus is simply a formula calculating an error. A limit is used to obtain a derivative. ‘The derivative is the slope of the tangent line at a given point. ‘The derivative is nothing but the rate of change. 40 CHAPTER | LET'S DIFFERENTIATE A FUNCTION!‘The derivative of flx) at x = is calculated by g(x) = Fa) (x ~ a) + fla) is then the approximate linear function of ftx). J's), which expresses the slope of the line tangent to fix) at the point (x. fly), is called the derivative of fix), because it is derived from ftx). Other than fx), the following symbols are also used to denote the derivative of y = fix) ay a a dx" dx" at) EXERCISES 1. We have function f(x) and linear function g(x) = 8x + 10. It is known, that the relative error of the two functions approaches 0 when x approaches 5. A. Obtain fi5). & Obtain s15). 2. Forte) =2°, obtain its derivative f10). exercises 412 LET'S LEARN DIFFERENTIATION TECHNIQUES!Criminal Charges Brought Against Megatrox Construction Contract Violates Antitrust Laws WOW! MEGATROX IS A HUGE COMPANY! 44 CHAPTER Z LETS LEARN DIFFERENTIATION TECHNIGUES!SUPPOSE YOU WANT TO WRITE A BIG ‘STORY SOMEDAY? torte ratepto \J BIG NEWS, NE ALSOWRITEN A }\ NOTHING TO BE PROUD ‘UNDERSTAND THAT YOU HAVE HIGH EXPECTATIONS FOR NEWSPAPER JOURNALISM, BUT THE BASICS ARE MOST NORKOWAKTSASCOORI a5.WRITE SIMPLY AND. CLEARLY-DON'T USE BIO WORDS OR JARGON. DON'T FORGET ABOUT THE READERS ON MAIN STREET. Solon Row aL We eeneene Schack PIOSHLIS smu FOU ORELE , DON'T PRETEND TO KNOW EVERYTHING! 46 CHAPTER 2 LET'S LEARN DIFFERENTIATION TECHNIGUES!COMPANIES AND STORES ARE ALWAYS TRYING TO SUPPLY CONSUMERS WITH BETTER MERCHANDISE AT LOWER PRICES. TT Th THE RESULT OF THEIR COMPETITION SHOULD. BE BETTER QUALITY AND LOWER PRICES, NOW, I WILL TELL YOU ABOUT A MOVING WALKWAY TO EXPLAIN WHY We MUST THINK OF THE ANTITRUST LAW IN TERMS OF CALCULUS. WELL, YOU KNOW THAT THE FEDERAL TRADE COMMISSION KEEPS AN BYE ON COMPANIES TO ‘SEE IF THEY DO ANYTHING THAT HINDERS FREE OF COURSE! COMPETITION, DON'T YOU? BUT IF SOME COMPANIES AGREE NOT TO COMPETE WITH EACH OTHER, OR SOMETHING ELSE HAPPENS TO HINDER COMPETITION, ‘CONSUMERS WILL BE GREATLY DISADVANTAGED. THE AIM OF THE FEDERAL TRADE COMMISSION 15 TO CONTROL SUCH ACTNITIES. WE'LL Discuss: THE SUM RULE OF DIFFERENTIATION. YOU SHOULD REMEMBER THIS BECAUSE IT IS USEFUL. NORIKO WANTS A ScOOPL aTHE SUM RULE OF DIFFERENTIATION LET'S LOOK ORMULA 2-1: INTO THIS BY THE SUM RULE OF DIFFERENTIATION APPROXIMATING ; AROUND x= a. For — h(x)=F(x)+9(x) (x)= f'(x) +9'(x) THAT Is, THE sae4e fin fiate-2)+ fia) 0 ee semanas a glW= glalx-a)+ g(a) e sermonmsins GIVEN THAT SINCE hia) = fu + ge, SUBSTITUTE © AND @ THIS EQUATION. hw~ktz-ai+f © 48 CHAPTER 2 LET'S LEARN DIFFERENTIATION TECHNIGUES!We ALSO KNOW THAT. hia) ~fialta-ar+ fray + glava-a) + g(a © 50 IF WE REARRANGE THE TERMS OF ©, EQUATION @ SAYS. THE COEFFICIENT OF, (e—a) WILL BE ky, k=fla) + gta! AND THE DIFFERENTIAL (COEFFICIENT EQUALS THE DERIVATIVE. £0, k=h{a)= Sia)+g'a). NOW, LET ME EXPLAIN ABOUT THE MOVING WALKWAY. YD RATHER NOT THINK ABOUT TT, BUT I GUESS TWILL. SUPPOSE FUTOSH! IS WALKING DOWN THE SIDEWALK. THE SUM RULE OF DIFFERENTIATION 40SUPPOSE THE DISTANCE HE WALKED SUPPOSE x MINUTES IN x MINUTES FROM LATER, HE 15 AT THE REFERENCE POINT POINT P. 016 Six) METERS. a MNUTES LATER, “| He IS AT POINT A. THAT'S RIGHT. BUT DOES TT MEAN ANYTHING? THIS MEANS THAT HE TRAVELED FROM A TO P IN (r= a) MINUTES, MR. SEKI, THE SUPPOSE THIS LEFT sD OF TRAVEL TIME THIS EQUATION 1S e- a) 15. (x)= F'(a)(x-a)+ $(a) DISTANCE TRAVELED EXTREMELY DINIDED BY TRAVEL ‘SHORT. TIME. £0, | THIS THE SPEED? THIS CAN BE CHANGED \_INTO.n StH) FA) x-a (a) S(@) REPRESENTS: FUTOSHI'S SPEED WHEN HE PASSES POINT A. 50 CHAPTER 2. LET'S LBARN DIFFERENTIATION TECHNIGUES!THAT MEANS THAT TO DIFFERENTIATE IS TO FIND THAT'S RIGHT. 50, IF THe SPEED WHEN fia) IG A < x) = flx) + glx), THEN FUNCTION EXPRESSING THE i Wi) = $8 + 9’) DISTANCE! 5 oO MEANS THE FOLLOWING. THIS TIME, LET Hit WALK ON A MOVING WALKWAY, LIKE YOU MIGHT SEE AT AN ‘AIRPORT. TRAVELS gtx) METERS Nx MINUTES THE MOVING WALKWAY MOVES fix) METERS IN x MINUTES, WHEN MEASURED ON THE WALKWAY, FUTOSHI TRAVELS (x) METERS IN x MINUTES. 50 THE TOTAL DISTANCE FUTOSHI TRAVELS IN x MINUTES BECOMES hx) =flx) + gle). ‘THE SUM RULE OF DIFFEREKTATION 51TT MEANS FUTOSHI'S TRAVEL THEN, WHAT DOES: SPEED, AS SEEN FROM he) =f) + 9'N) ‘SOMEONE NOT ON THE WALKWAY, (MEAN? 15 THE SUM OF HIS SPEED ON ‘THE WALKWAY AND THE SPEED. OF THE WALKWAY ITSELF, BE PATIENT. BUT, I'S NOT 50 FORA LITLE ‘SURPRISING, 15, WHILE LONGER, IT? DOES THIS GRASSHOPPER, HAVE ANYTHING TOLD YOU THAT TO DO WITH THE THE BASICS ARE ANTITRUST LAW? IMPORTANT. THE NEXT RULE IS ALSO FUNDAMENTAL, $0 REMEMBER THIS ‘ONE, TOO. 52 CHAPTER 2 LET'S LEARN DIFFERENTATION TECHNIQUES!THE PRODUCT RULE OF DIFFERENTIATION pe et oe eer (ses For n(x)=S(x)9() h(x) = S(x)9(x) + F(x) 9'(x) of the products with only one function differentiated. S (x)= S'(a)(x-a)+ (a) 9(x)=9'(a)(x-a)+9(a) (x)= 5 (x) 9(x) =k(x-a) +t n(x) #{F(a)(x-a) + $(a)}x{9'(a)(x-a)+9(a)} h(x) = s'(a)9'(a)(x-a)' + F(a) 9'(a)(x- a) + F'(a)(x-a) g(a) + F(a) 9(a) (x -a) 1S A SMALL CHANGE. THAT MEANS: THROW THAT TERM OUT, h(x) ={F'(a)9(a)+ F(a) 9'(a)}(x-@) + F(a) g(a) k= S'(a)g(a) + f(a)a’(a) THE PRODUCT RULE OF DIFFERENTIATION 53NOW, I WILL USE spent seer ven ea Wace Beara, ten NOT BE ALLOWED. HOW Do you SOLVE A SOCIAL PROBLEM USING DIFFERENTIATION? LET'S LOOK AT THE WORLD INA MORE BUSINESSLIKE MANNER, THAT'S RIGHT.” COMPANIES IN A PERFECTLY COMPETITIVE MARKET ACCEPT THE COMMODITY PRICE DETERMINED BY THE MARKET AND CONTINUE: A MARKET WHERE MARY COMPANIES SUPPLY. PRODUCTS THAT CANNOT BE DISCRIMINATED BETWEEN 15 CALLED "A Baeteny cowpemne iO paopies sib SubPLy MARKET.” THEIR PRODUCT AS LONG H AS THEY MAKE PROFITS. —_ cient oe PERFECTLY EXAMPLE? | COMPETITIVE MARKET Ler’s See... VIDEO RENTAL ‘sHoPs? ™ DX REALITY, THERE ARE USUALLY BIO-NANE BRANDS FOR ANY COMMODITY. 54 cHaPteR 2 THERE ARE FAMOUS CHAIN SHOPS INTHE VIDEO RENTAL MARKET-NO MARKET CAN BE APERFECTLY COMPETITUE ONE, 50 THIS 15 A FICTIONS, IDEAL SITUATIONIF THE COST OF PRODUCING ONE MORE UNIT IS 10,000, THE COMPANY WILL SURELY INCREASE PRODUCTION, BECAUSE IT WILL MAKE ‘MORE PROFIT. ‘SUPPOSE, FOR EXAMPLE, A COMPANY PRODUCING CD PLAYERS WHOSE MARKET PRICE IS ¥1Z,000 PER UNIT CONSIDERS WHETHER OR NOT IT WILL INCREASE PRODUCTION VOLUME. SINCE MANY (OTHER COMPANIES PRODUCE THE SAME KIND OF PRODUCT, THE ‘COMPANY BELIEVES: THAT ITS INCREASE IN PRODUCTION WILL CAUSE THE PRICE TO DECREASE. 50 THE COMPANY WILL CONSIDER MAKING ADDTIONAL UNITS. BUT THE COST OF MAKING ONE MORE UNIT (CHANGES, AND THE COMPANY'S PRODUCTION EFFICIENCY WILL CHANGE. EVENTUALLY, THe COST (OF MAKING ONE MORE UNIT WILL REACH THE MARKET PRICE (OF ¥12,000. AT THAT POINT, AN INCREASE IN PRODUCTION WOULD NOT 8 WORTH THE COST. IN SHORT, THE MARKET STABILIZES WHEN THE MARKET PRICE OF THE UNIT EQUALS THE COST OF PRODUCING ‘ANOTHER UNIT. ON THE OTHER HAND, THE WHEN YOU LooK STORY 1S DIFFERENT INA IAT THE MARKET MONOPOLY MARKET, WHERE AS A WHOLE, AN JONLY ONE COMPANY SUPPLIES “J INCREASE IN SUPPLY ‘A PARTICULAR PRODUCT. THEN 7} WILL CAUSE THE JUST ONE COMPARY 15 THE PRICE TO GO POWN| ee eels > AND DEMAND. EE monoroty \HEHH) marker / THe PRODUCT RULE OF DIFFERENTIATION 95NOW, LET'S ASSUME WE KNOW THAT THE PRICE THAT ALLOWS: BY THE WAY, p'tx), WHICH EXPRESSES THE CHANGE IN PRICE, IS NEGATIVE BECAUSE THE UNIT'S PRICE DECREASES IF THE COMPANY TO SELL EVERY UNIT SUPPLIED IN QUANTITY x IS pix, ‘A PUNCTION OF x. THAT'S RIGHT. THE COMPANY'S. REVENUE FROM THIS PRODUCT IS GIVEN BY THis... FORMULA 2-3: ‘THE COMPANY'S REVENUE Since R(x)~ R’(a)(x-a)+R(a) we know that R(x)-R(a)=R(a)(x-a) af uence PRODUCTION Vouume lance REVENUE I GET IT! THE COMPANY NEEDS TO CALCULATE THIS TO DECIDE WHETHER TO INCREASE PRODUCTION, WHILE COMPARING IT AGAINST THE COSTS OF PRODUCING THE UNITS. 56 CHAPTER 2. LET'S LEARN DIFFERENTIATION TECHNIQUES! x IS INCREASED. Revenue = R(x) = price x quantity = p(x) xx THIS SHOWS US THAT THE ADDITIONAL REVENUE FROM AN INCREASE IN PRODUCTION IS Ra) PER UNIT. YOU'RE RIGHT. SINCE RUx) = pla) xx, REMEMBER THAT PRODUCT RULE OF DIFFERENTIATION. THINK REMEMBER...We GET Ria) = pa) xa+ pla) x1 RIGHT. PRODUCTION SHOULD BE STOPPED AT THE “THE DERIVATNE OF x 161 (GEE PAGE 40 FOR MORE ON DIFFERENTIATING LINEAR FUNCTIONS), PRODUCTION WILL BE STOPPED WHEN p'(a) x a+ pla) = (OF PRODUCTION: We KNOW ‘THAT THE FIRST TERM IS NEGATIVE, SO THE MARKET PRICE pla) IS GREATER ‘THAN THE COST. BUT THE PRICE IS ACTUALLY GREATER THAN THE COST OF PRODUCING AN ADDITIONAL UNIT WHEN ‘A MONOPOLISTIC COMPANY STOPS PRODUCTION. THAT'S UNDUE PRICE-FIXING, YOU ARE RIGHT, BUT YOU ‘SHOULD TAKE A CLOSER LOOK. COMPANIES DO- THIS NOT BECAUSE OF MALICIOUS MOTIVES BUT BASED ON A RATIONAL. JUDGMENT. LOOK AT THE EXPRESSION Py THE PRODUCT RULE OF DIFFERENTIATION 57Sales increase (per unit) when production is increased a little more: R(a)=p'(a)a+ p(a) ‘The two terms in the last expression mean the following: (a) represents the revenue from selling a units P'la)a = Rate of price decrease x Amount of production = A heavy loss due to price decrease influencing all units WHAT DO YOU WHAT DO THE MONOPOLY THINK, Noriko? / \ T THINK? 1] s1OPS PRODUCTION, CONSIDERING BOTH HOW MUCH IT OBTAINS BY SELLING ONE MORE UNIT AND HOW MUCH LOSs IT SUFFERS DUE TO A PRICE DECREASE. IF 60, IT 1S NOT DOING. BUT, FOR CONSUMERS: ‘A "BAD" THING BUT Is AND SOCIETY, THE aaa eee \CCORDANCE WITH A CAPTIALIST PRINCIPLE PRICES, WHICH IS NOT OF PROFIT-SEEKING, DESIRABLE. THAT'S THEREFORE, ACCUSING WHY MONOPOLIES ARE ‘THE COMPANY OF BEING \\ PROHIBITED BY LAW. MORALLY WRONG IS OF NO USE, 58 CHAPTER 2 LET'S LEARN DIFFERENTIATION TECHNIGUES!MR. SEKI, THAT'S GREAT ALL OF SOCIETY'S PROBLEMS CAN BE SOLVED WITH DIFFERENTIATION, WHAT ABOUT LOVE? HOW DO YOU SOLVE THE PRODUCT RULE OF DIFFERENTIATION 52THE NEWSPAPER WANTS TO ASK YOU ‘A FEW QUESTIONS: BOUT 60 CHAPTER 2 LET'S LEARN DIFFERENTIATION TECHNIGUES!THEY WANT TO KNOW MORE ABOUT YOUR SOURCES AND Yes..1 UNDERSTAND, THANK YOU FOR CALLING ME. PLL GET EVERYTHING TOGETHER, WHAT'S THE MATTER? YOU DONT LOOK ‘50 GOOD. Ma. SER GETS ACALL 61DIFFERENTIATING POLYNOMIALS: AS A WRAP-UP, 3 MemoRize THE FORMULAS FoR wy POLNOMALS He "ALS. ely ee DIFFERENTIATION ‘OF ANY POUNOMIAL CAN Be PeRPORMED By COMBINING, “THREE FORMULAS. FORMULA 2-4: THE DERIVATIVE OF AN nTH-DEGREE FUNCTION The derivative of f(x) =x" is h'(x)=nx"* How do we get this general rule? We use the product rule of differentiation repeatedly. For h(x)= 27, since h(x) = «x x,h’(x) =xx1+ Lex =2x ‘The formula is correct in this case. ‘THIS RESULT 16 USED For h(x) = x*, since R(x) =x" «x, W(x) =(x*) xx +a" x(x) =(2x)x +x x ‘The formula is correct in this case, too. For h(x) =x*, since h(x) =x° xx, h’(x)=(x") xx +x" x(x) =x? xxtax* x1 = 4x" Again, the formula is correct. This continues forever. Any polynomial can be differentiated by combining the three formulas! FORMULA 2-5: THE DIFFERENTIATION FORMULAS OF SUM RULE, CONSTANT MULTIPLICATION, AND x" © Sum rule: {5(x)+9(x)} =s'(x)+9'x) © Power rule (x © Constant multiplication: {«f(x))'=«s'(x) Let's see it in action! Differentiate h(x) =x? + 2x" +5x +3 rule @ W(x) = {x2 +282 +54 +3 = (0°) +(20°) + (6x) + (8) PY 42(xt) +5(x) = 3x" 4 2(2x)+8x1=3x" 44x45 rule Tule ® 62 CHAPTERZ LET'S LEARN DIFFERENTIATION TECHNIQUES!basis aii i re ) P7FINDING MAXIMA AND MINIMA ill ¢ ai a 5 hf x f Ys H ees ey S oe, i fi ¢ Soom : : LIKE A Minimum x Bo ice eae] J BS Se Maxima and minima are where a function changes from a decrease to an increase or vice versa. Thus they are important for examining the properties ofa function, Since a maximum or minimum is often the absolute maximum or minimum, respectively, it is an important point for obtaining an optimum solution, THEOREM 2-1: THE CONDITIONS FOR EXTREMA Ify =J(x) has a maximum or minimum at x= a, then f(a) = 0. This means that we can find maxima or minima by finding values of a that satisfy f'(a) = 0. These values are also called extrema. War's THAT? EHATE ROLLER COASTERS... ctack ackery 64 CHAPTER 2 LET'S LEARN DIFFERENTIATION TECHNIQUES!Assume f(a) > 0. Since fix) = f(a) (x - a) +fla) near x= a, ‘JS'(@) > O means that the approximate linear function is increasing at x= a. ‘Thus, so is fix) In other words, the roller coaster is ascending, and it is not at the top or at the bottom. Similarly, y = flx) is descending when f(a) < 0, and it is not at the top or the bottom, either. Ify =flx) is ascending or descending when Ja) > 0 or f(a) <0, respectively, ‘we can only have f'(a) = 0 at the top or bottom. In fact, the approximate linear function y =J'(a) (x - a) + fla) = 0 x (x - a) + fla) is a horizontal constant function when f'(a) = 0, which fits our under- standing of maxima and minima. (afta) \ / 7 \ Sa) =0 (a) =0 z ta. fla) THEOREM 2-2: THE CRITERIA FOR INCREASING AND DECREASING =Jix) is increasing around x = a when f(a) > 0. y =Jlx) is decreasing around x = a when fa) <0. FINDING MAXIMA BND MINIMA 65bry Seer a OH, $0 YOU UNDERSTAND! WHAT? YOU HAVE: ANYTHING NEW TO SAY? ALL YOU SAY 1S DIFFERENTIATION, ‘DIFFERENTIATION. NO, THANK YOU. I DON'T IT'S BECAUSE WANT TO DRINK TOO OF THAT CALL, MUCH TONIGHT. ISN'T IT? WHAT 66 CHAPTER 2 LET'S LEARN DIFFERENTIATION TECINIGUES!DELICIOUS! DRAFT BEER IS THE BEST AND RELATIVELY LARGE ONES THAT QUICKLY BECOME LARGER, RISE UP TO THE ‘SURFACE, AND POP THERE. NOW, EXPLAIN WHY THIS HAPPENS! ‘SMALL ONES THAT BECOME EVEN SMALLER AND <<< FINALLY DISAPPEAR... FINDING MAXIMA AND MINIMA 67SINCE CARBON DIOXIDE IN CARBONATED DRINKS, SUCH AS BEER, IS SUPERSATURATED, IT 15 ‘MORE STABLE AS A GAS THAN WHEN IT IS DISSOLVED IN FLUID. G0, THE ENERGY OF A BUBBLE DECREASES IN PROPORTION TO 73 VOLUME (Sar*, WITH r BEING THE 3° RADU). ON THE OTHER HAND, SURFACE GAS (BUBBLE) TENSION ACTS ON THE BOUNDARY SURFACE BETWEEN THE BUBBLE AND THE FLUID, TRYING TO REDUCE THE SURFACE AREA. oe - aCe a Tmee OTe THEREFORE, THE ENERGY OF ° ‘THE BUBBLE DUE TO THIS FORCE INCREASES IN PROPORTION TO THE SURFACE AREA, 4nr*, CONSIDERING THESE TWO EFFECTS, THE ENERGY E(r) OF A BUBBLE OF RADIUS r CAN BE EXPRESSED 68 CHAPTER 2 LEY'S LEARN DIFFERENTIATION TECHNIGUES!TO SIMPLIFY THE PROBLEM, LET'S ASSUME a AND b ARE 1 AND CHANGE THE VALUE OF r $0 THAT E(t) =—1° + 3r° THAT 1S ENOUGH TO SEE THE GENERAL SHAPE OF E(r), THE BUBBLE TRIES TO. REDUCE ITS ENERGY AS MUCH AS POSSIBLE. IF WE FIND OUT HOW Eir) BEHAVES TO REDUCE ITSELF, We WILL SOLVE FIRST, LET'S NE FIND THE E'(r)=(-7°) +(ar°) EXTREMUM. 2 =-3"? +6r -Br(r-2) WHEN r= 2, E'(r) = 0, FOR O 0), THE FUNCTION IS INCREASING, AND FOR 2 Reauiges ALoT OF TIME TO WRITE DOWN, I WILL SHOW YOU ITS ‘SYMBOL. oe 40 CHAPTERS LETS INTEGRATE A FUNCTION!USING THE FUNDAMENTAL THEOREM OF CALCULUS Pluk t+ Plata Di) ont Plasite- 25) BUT, WHAT 15 AP THE ABOVE PxeaSN 2 pide A= Le, Ls... ‘CAN BE WRITTEN) ( IN THIS WAY. of, SIMPLE! A (DELTA IS A GREEK ? LETTER. THE SYMBOL IS Sraaaercs USED To EXPRESS THE ‘AMOUNT OF CHANGE. THIS Ax EXPRESSES THE DISTANCE TO THE NEXT POINT. IN OTHER WORDS, TT 16, FOR EXAMPLE, (4, =o) OR (0 - 1) USING THE FUNDAMENTAL THEOREM OF CALCULUS a}EXPRESSES THE OPERATION “GUM UP FROM x, = 0 TOx,=9." IT MEANS TO SUM UP THE VALUE OF p ATO TIMES ] CTHE DISTANCE FROM x TO THE NEXT POINT) ZA THE NEXT ONE IS THE SYMBOL TO SIMPLIFY THIS EQUATION FURTHER. SINCE THE EQUATION IS THE SUM FOR A FINITE at Be a NUMBER OF STEPS, ME MM | =| We MAKE THE SYMBOL ROUND WHEN WE HAVE: AN INFINITE NUMBER OF 2 CHAPTERS LET'S INTEGRATE A FUNCTION! NOW NORIKO, WHAT DOES D_ plxjae YES, IT MEANS: THE EQUATION WE SAW BEFORE AT THE BOTTOM OF PAGE 89,I EXPAND > TO. MAKE |, AND IF WE KNOW p(x) 16 THE DERIVATIVE Daca @ > THIS 1S THE FUNDAMENTAL THEOREM OF CALCULUS! EXPRESSION © MEANS THE SUM WHEN THE INTERVAL IS MADE INFINITELY SMALL, AND TT EXPRESSES THE AREA BETWEEN THE GRAPH ON THE LEFT AND THE X-AXIS. THIS 1S CALLED A DEFINITE INTEGRAL. [i p(x)ax=q(b)-a(a) We HAVE CALCULATED THE SUM EXTREMELY EASILY IN THIS WAY, HAVENT WEP p(x)=fp(xjac= plae=a(e)-ate) | We must find q(x) that satisties q(x) = p(x)a. > USING THE FUNDAMENTAL THEOREM OF CALCULUS 45A STRICT EXPLANATION OF STEP 5 In the explanation given before (page 89), we used, as the basic expression, q(x,)~a(x)= B(%)(&, ~¥o)e “crude” expression which roughly imitates the exact expression. For those who think this is a sloppy expla- nation, we will explain more carefully here. Using the ‘mean value theorem, we can reproduce the same result. We first find q(x) that satisfies g's) = peo. We place points x9 (= 4), Xa, » X, (= b) on the xaxis, We then find point xp, that exists between x, and x, and satis- fies a(x)~4a(%) = 4 (&)(% —% Xe +(x, )~ 4%, Theexistenceofsuchapoint, = gt gg ee is guaranteed by the mean value id Beccregercmeieicaveccae theorem, Similarty, we find x, kk, =a) kx, =) between x, and'x, and get a(x) -a(s)=41(%) 6 — ‘Areas of these steps Repeating this operation, we get 2 4()-a(x)= Fal —*) = Plas) | B a(m)-alx)= ¥(a)05 5) Dosa) | E 4(%)-al)= 1%) 8%) osx) 1 (Xn-in)(%s ~Xu-a) = P(Xan ) (Xs a(x, )-a(%.) a(b)-a(a) +—! Equal ‘This corresponds to the diagram in step 5. ‘Approximate area | intinitey fine sections Always equal [— exact area 94 CHAPTER 3. LETS SKTEGRATE A FUNCTION!USING INTEGRAL FORMULAS FORMULA 3-1: THE INTEGRAL FORMULAS © Ps lxjaes [Es (x)ae= [2 F(x)ax The intervals of definite integrals of the same function can be Joined. © [{r(x)+9(x)}ae= [2 (x)ax+ [Po (x)ax A definite integral of a sum can be divided into the sum of defi- nite integrals. © PPas(x)de=af" s(xjax The multiplicative constant within a definite integral can be moved outside the integral. Expressions © through @ can be understood intuitively if we draw their figures. ° in i casa aD Aue aaa ° ‘Area for 9 e atx) Area tors | F nwa a B a a . ets te) C= b reais ‘$29 ~ multiplied ~ = | USING THE FUNDAMENTAL THEOREM OF CALCULUS 45WHEW! We: ARE ALL DONE. FUTOSH, HELP YOURSELF TO SOME SHOCHU. IVE JUST REMEMBERED ‘A TASK FOR YOU. WILL YOU GO TO THE REFERENCE ROOM? 96 CHAPTER 3 LET'S INTEGRATE A FUNCTION!NORIKO, I REMEMBER THAT ABOUT A YEAR AGO, A GROUP ( | WHY Do THey (OF RESEARCHERS AT SANDA KEEP BRUSHING ENGINEERING COLLEGE ‘ALSO ANALYZED WIND. CHARACTERISTICS AND USED THEIR RESULTS TO DESIGN BUILDINGS. WILL YOU FIND OUT HOW THEIR RESEARCH HAS PROGRESSED SINCE THEN? KAKERU SEKI. weer is ir THIS 15 AN ARTICLE More ‘THE REFERENCE ROOM 47BURNHAM... THEY'RE ONE OF THE SPONSORS OF THE ASAGAKE TIMES. aut aye wea n THAT MUST BE WHY HE WAS TRANSFERRED TO ‘THIS BRANCH OFFICE. OUR BIGGEST ADVERTISER. 26 CHAPTER 3 LET'S INTEGRATE A FUNCTION!HAVE YOU FOUND ‘ANYTHING? NO, WELL...AH, THEY PROPOSED INTERESTING IDEAS, OH, THAT'S GOOD. 60, WHAT KIND OF ‘THE REFERENCE ROOM 99FORGET ABOUT CALLING! Go See THEM FOR AN INTERVIEW! AND AS PUNISHMENT, YES, SIR! FIND OUT IF THEIR IM ON MY WAY. CAN BE WRITTEN USING EQUATIONS! 100 CHAPTER 3 LET'S INTEGRATE A FUNCTION![R= APPLYING THE FUNDAMENTAL THt wep EXACTLY! IN ECONOMICS, THE INTERSECTION OF ‘THE SUPPLY AND DEMAND CURVES IS DETERMINE THE PRICE AND QUANTITY AT WHICH COMPANIES PRODUCE AND uae eer, IN TRUTH, SOCIETY 15 BEST SERVED IF TRADE MATCHES THESE IDEAL CONDITIONS. APPLYING THE FUNDAMENTAL THEOREM 101‘SUPPLY CURVE FIRST, LET'S CONSIDER HOW COMPANIES MAXIMIZE PROFIT IN A PERFECTLY COMPETITIVE MARKET. WE'LL TRY TO DERINE A SUPPLY CURVE FIRST. ‘The profit P(x) when x units of a commodity are produced is given by the fol- lowing function: (Profit) = Price) x (Production Quantity) ~ (Cost) = px Ctx) where C(x) is the cost of production. Let's assume the x value that maximizes the profit P(x) is the quantity of production s. Accompany wants to maximize its profits. Recall that to find a function's extrema, we take the derivative and set it to zero. This means that the com- pany's maximum profit occurs when P(s)=p-C(s)=0 P (Price) THE FUNCTION p = C'(s) OBTAINED ABOVE 1S CALLED THE SUPPLY CURVE! s (Optimum production volume by companies) Price p, corresponds to point A on the function, which leads us to opti- mum production volume s, 102 CHAPTER 3 LET'S INTEGRATE A FUNCTION!The rectangle bounded by these four points (p,. A, s,, and the origin) equals the price multiplied by the production quantity. This should be the companies’ gross profits, before subtracting their costs of production. But look, the area © of this graph corresponds to the companies’ production costs, and we can obtain it using an integral. [he(s)as~c(s,)-C(o)=c(s,) costs NY NS Weasea the Fondant oratce ‘This means we can easily find the companies’ net profit, which is repre- sented by area @ in the graph, or the area of the rectangle minus area @. DEMAND CURVE Next, let's consider the maximum benefit for consumers. When consumers purchase x units of a commodity, the benefit B(x) for them is given by the equation: B(x) = Total Value of Consumption - (Price x Quantity) = ls) = pe where u(x) is a function describing the value of the commodity for all ‘Consumers will purchase the most of this commodity when B(x) is maximized. Ifwe set the consumption value to t when the derivative of Blx) = get the following equation” Bi(t)=w'(t)-p=0 ‘THE FUNCTION p = w'(t) OBTAINED HERE 1S, CALLED THE DEMAND CURVE. + Again, you can see we're looking for extrema (where Blt) = 0}, as consumers want to mi mize thelr benefits APPLYING THE FUNDAMENTAL THEOREM 103Pp, ¢ (Optimum consumption) o| ft, So let's consider the area of the rectangle labeled ©, above, which corre- sponds to the price multiplied by the product consumption. In other words, this is the total amount consumers pay for a product. The total area of © and @ can be obtained using integration. fiw (o)dt=u(,)-u(0)=u(¢,) = Total value of consumption Se To simply, 0} = 0. Ifyou simply subtract the value of the rectangle @ from the integral from 0 to f,, you can find the area of ®, the benefit to consumers. THE BENEFIT FOR THE CONSUMERS © 15 THE TOTAL, VALUE OF CONSUMPTION MINUS | | YES, THAT'S IT. NOW LET'S i 2 LOOK AT THe SUPPLY THE AMOUNT THEY PAID ©, RIGHT? Leona ae 5 COMBINED TOGETHER, 104 CHAPTER 3. LET'S INTEGRATE A FUNCTION!P (Price) Supply curve WE CAN SAY THAT THE COMPANIES’ PROFIT PLUS THe BENEFIT FOR CONSUMERS EQUALS THE OVERALL BENEFIT FOR SOCIETY, AS ILLUSTRATED BY THE SHADED AREA ON THE RIGHT. Demand curve Quantity BUT WHAT HAPPENS IF TRADE DOES NOT HAPPEN AT THE PRICE ‘AND QUANTITY DETERMINED BY THE INTERSECTION PONT E? Loss of ‘THE OVERALL BENEFIT TO SOCIETY eee 1S REDUCED BY THE AMOUNT CORRESPONDING TO THE EMPTY AREA society IN THE FIGURE. yes, TWILL REPORT MY STORIES USING CALCULUS, TOO. ALSO THINK VELOCITY AND FALLING BODIES ARE GOOD TOPICS TO WRITE, IM GOING TO LOOK INTO THEM! ‘APPLYING THE FUNDAMENTAL THEOREM 105¥50 THE CALCULUS NEWS-GAZETTE IVol. 1 The Integral of Velocity Proven to Be Distance! The integral of velocity = difference in position = distance traveled If we understand this formula, it’s said that we can correctly calculate the distance traveled for objects whose veloc- ity changes constantly. But is that true? Our promising freshman reporter Noriko Hikima closes in on the truth of this ma ter in her hard-hitting report. Figure 1: This graph represents Futoshi's distance traveled over time. He moves to point yj, Yo. Ys as time progresses to x,, Xs. Xy [A] vetocity Figure 2 eee Sanda-Cho—Some readers will recall our earlier example describing Futoshi walk- ing on a moving walkway. Others have Hikely deliberately blocked his sweaty image from their minds. But you almost certainly remember that the derivative of ‘the distance is the speed. © y-F(x) © flu(xjax Equation © expresses the position of the monstrous, sweating Futoshi. In other words, after x seconds he has lumbered a total distance of y. (0)- F(a) Integral of Velocity ference in Position The derivative F(x) of expression @ is the “instantaneous velocity” at x sec onds. If we rewrite Fx) as v(x), using v for velocity, the Fundamental Theorem of Ci culus can be used to obtain equation Look at the graph of v(x) in Figure 2-A— Futoshi’s velocity over time. The shaded part of the graph is equal to the integral— equation ©. But also look at Figure 2-B, which shows the distance Futoshi has traveled over time. If we look at Figures 2-A and 2B side by side, we see that the integral of the velocity is equal to the difference in position (or distance)! Notice how the two graphs match— when Futoshi velocity is posi- tive, his dis Distance tance increases, and vice versa.‘The Caleulus News-Gazette Free Fall from Tokyo Tower How Many Seconds to the Ground? It’s easy to take things for granted— consider gravity. If you drop an object from your hand, it naturally falls to the ground. We can say that this is a motion that changes every second—it is accelerating due to the Earth's gravitational pull. Thi motion can be easily described using calculus. But let's consider a bigger drop—all the ‘way from the top of Tokyo Tower—and find out, "How many seconds does it take an ‘object to reach the ground?” Pay no atten- tion to Futoshi's remark, "Why don't you go to the top of Tokyo Tower with a stopwatch and find out for yourself?" The increase in velocity when an object is in free fall is called gravitational acceleration, or 9.8 mis*. In other words, this means’ that an object's velocity increases by 9.8 m/s every second. Why is this the rate of acceleration? Well, let's just assume the scientists are right for today. Expression © gives the distance the object falls in T seconds, Since the integral of the velocity is the difference in position (or the distance the object travels), equ: tion @ can be derived. Look at Figure 3— we've calculated the area by taking half of the product of the x and y values—in this case, % x 9.8t x f. And we know that the height of Tokyo Tower is 333 m. The square root of (833 / 4.9) equals about 8.2, so an object takes about 8.2 seconds to reach the ground. (We've neglected air resistance here for convenience.) ‘Section AL Velocity ox) = 9.8 ‘Area ofthe velocity 88x t= 396 Distance a fatten ¢The Caleulus News-Gazette The Fundamental Theorem of Calculus Applies to Dice, Too You probably remem- | ¥,| ber playing games with dice as a child. Since ancient times, these hexahedrons have been rolled around the world, not only in games, but, also for fortune telling and gambling Mathematically, you can say that dice are the world's smallest random-number generator. Dice are wonder- ful, Now we'll cast them for caleulust A die can show a 1, 2,3, 4, 8, or 6—the probabil. ity of any one number Is 1 in 6. This can be shown with a histogram (Figure 4), with their numbers on the x-axis and the probability on the yeaxis, ‘This can be expressed by ‘equation ©. of fix) = Probabil- ity of rolling x. This becomes equation © when we try to predict a single result—for example, aroll of 4 fee) © J(s)= Poti oroting © s(a)-2 Prosi orsoing 4 Now let's take a look at Figure 5, which describes a distribution function. First, start at 1 on the xaxis. Since no number less than 1 exists on a die, the probability in this region is 0. Atx= 1, the graph jumps to 1/6, because the probability of rolling a ‘number less than or equal to 1 Is 1 in 6. You can also see that the probability of rolling a number equal to or greater than 1 and less than 2is 1/6 as well. This should make intuitive sense. At, the probability jumps up to 2/6, which means the probability for rolling a number equal to or less than 2 is 2/6. Since this probability remains until The Die Is Cast!!! yay 1234656 Figure 4: Density function Density function 123456 SSS SESS esses EESSSEEESSS| Figure 6: Derivative of distribution funetion F(x) lensity function flx) Section AL Figure 5: Distribution function Distribution function Fix) 1 % 123456 S(3)+5(4)+5(5)=F(5)-F(2) right below 8, the probability of numbers ess than 3 is 2/6 © fy(x)ae=F (b)-F(a) Probability of rolling x where asx In the same way, we can find that the probability of rolling a 6 or any number ‘smaller than 6 (that is, any number on the die) is 1. After all, a die cannot stand on fone of its corners. Now let's loole at the probability of rolling numbers greater than 2 and equal to or less than 5. The equation in Figure 6 explains this relationship. If we look at equation ©, we see that it describes what we know—"A definite inte- gral of a differentiated function = The dif- ference in the original function.” Thi nothing but the Fundamental Theorem of Calculust How wonderful dice are.ONLY 15 MINUTES C\ TO GET THERE! A LHAVE To REPORT fovsr ON THE SANDA-CHO § SUMMER FESTIVAL, Jo”REVIEW OF THE FUNDAMENTAL THEOREM OF CALCULUS: When the derivative of F(x) is flx), that is, if fx) = Fx) [LS (x)ae=F(b)-F(a) This can also be written as [LF '(x)dx = F(b)- F(a) These expressions mean the following. (Differentiated function) dx = Difference of the original function between b and a Italso means graphically that ‘Area surrounded by the differentiated function ‘Change in the original and the x-axis, between x= a and x=b function from a to b y=JS(s)=F(x) y it =F (x) Fi(b) a 0 ( Fundamental Fa teomn | Y a of Calculus fee - aah Difference in the original function NO CHAPTERS. LET'S INTEGRATE A FUNCTION!FORMULA OF THE SUBSTITUTION RULE OF INTEGRATION When a function of y is substituted for variable x as x = gly). how do we express S= [Ps (x)ax a definite integral with respect to x, as a definite integral with respect to y? First, we express the definite integral in terms of a stepwise function approximately as SD Fl) a-ak) (= Gx, =b) ‘Transforming variable x as x= gly), we set Yo = OY ¥ “8 so that a= 9(a),x, = 9(4).%1 = 9 (Ya) Note here that using an approximate linear function of G (Yarn) ~ (Ye) = 9 (Ye) (Yaar ~ Ye) Substituting these expressions in S, we get s- > FN D_ F(9(4e)) 9 (Ye) Yee The last expression is an approximation of [s(ow)a'uyay Therefore, by making the divisions infinitely small, we obtain the follow- ing formula, FORMULA 3-2: THE SUBSTITUTION RULE OF INTEGRATION fs (e)ax =f! s(9(y)) 9 (vay FORMULA OF THE SUBSTITUTION RULE OF INTEGRATION 11EXAMPLE: Calculate: [roe +)'ae ‘We first substitute the variable so that y = 2x + 1, or x= gly)= Since y = 2x + 1, ifwe take the derivative of both sides, we get dy = 2dx. Then we get dx. ‘Since we now integrate with respect to y, the new interval of integra- tion is obtained from 0 = g(1) and 1 = g(3) to be 1-3. f10(2x+1)' ae = ffaout Say =['Sy'ay =3° -1° = 242 THE POWER RULE OF INTEGRATION In the example above we remembered that 5y* is the derivative of y° to finish the problem. Since we know that if F(x) = x", then F'(x) = fix) = nx"~", we should be able to find a general rule for finding F(x) when (x) = x”. ‘We know that F{x) should have x" in it, but what about that coef ficient? We don't have a coefficient in our derivative, so we'll need to start with one. When we take the derivative, the coefficient will be (n+ 1), 80 it follows that 1 / (n + 1) will cancel it out. That means that the general rule for finding the antiderivative F(x) of fix) = x" is Le fot ot at xatl 2 x nel F(x $n other words, when. detinite integral. - y= 1, and when x= 1, y 3. We then use that as the range of our 2 CHAPTER 3 LET'S INTEGRATE A FUNCTION!EXERCISES 1 Caleulate the definite integrals given below, © [faxae ee © fixe (tex) aes fix- (Lex) ae 2, Answer the following questions. ‘Write an expression of the definite integral which calculates the area surrounded by the graph of y = flx) = x° ~ Sx and the xaxis. ®& Calculate the area given by this expression. EXERCISES 134 LET'S LEARN INTEGRATION TECHNIQUES!USING TRIGONOMETRIC FUNCTIONS: rena mn O« eay 1-9 PN aaa Zu f i : ; WHEW! I MADE IT, JUST IN TIME. ¥ we SS I WANNA PUT, (ON A YURATA’ = YUKATA IS TRADMIONAL JAPANESE SUMNER WEAR. pee noriko, you're Here. |}, \ /EAH..SINCE I HAVE A IT WAS NICE OF YOUTO|| CELL PHONE, 1 CAN'T CALL AND LET Me KNOW REALLY GET AWAY YOU MIGHT BE RUNNING FROM YOU. LATE. % N6 CHAPTER LETS LEARN INTEGRATION TECHNIQUES!WHEN T WAS A CUB REPORTER, THERE WASN'T SUCH A CONVENIENCE. LOFTEN HAD TO. USE A PAY PHONE TO SEND IN MY REPORT WHEN 1 WAS ON DEADLINE. ALL SORTS OF OTHER WAVES OccuR IN NATURE, TOO. LEAD My REPORT WORD BY WORD OVER ‘THE PHONE TO MY ! ASSISTANT. WE DONT HAVE TO DO THAT ANYMORE, THANKS TO RADIO WAVES. OCEAN WAVES, EARTHQUAKES, SOUND WAVES... ‘AND LIGHT. USING TRIGONOMETRIC FUNCTIONS. 17THOSE WAVES CAN BE DESCRIBED WITH FUNCTIONS, FOR EXAMPLE, WITH THE ‘COSINE OF THETA (os @. DID YOU KNOW THAT? UH, I HAVE TO GO BACK TO WORK. NORIKO! INCIDENTALLY, IF YOU CUT OUT A SLEEVE OF ‘A BLOUSE, THE CUT END IS A GRAPH OF NB CHAPTER 4 LET'S LEARN INTEGRATION TECHNIGUES!J LOOK AT THE NORIKO, TAKE A DANCERS, THIS IS A PICTURE OF THAT! |GOOP OPPORTUNITY. IT'S cos 0. WE CAN STUDY THE APPLICATION OF FUNCTIONS. YOU AND YOUR FUNCTIONSII! TOGETHER WHILE AUNT OF MEASUREMENT (OH, SHOOT! 7M. FOR ANGLES TAKING NOTES CALLED A OUT OF HABIT. RADIAN, OF RADIUS 1 WITH FOR A CIRCLE WITH 1S CENTER AT (0, 0). RADIUS = 1, THE Baer tute en or he ener Seucane ald THE ANGLE 0 IN a eaaie (OF THE CIRCLE, RADIANS! CORRESPONDING TO, THE ANGLE 6. USING TREONOMETRIC FUNCTIONS. 1K¢BECAUSE THE TOTAL FROM NOW ON, CIRCUMFERENCE OF THIS |) Purnene ICIRCLE 15 2x, WE KNOW THAT P ANS 90 DEGREES = | RADIANS AND 160 DEGREES= x RADIANS. A RADIAN IS ABOUT EQUAL TO 572958 DEGREES. "AND WE CAN EXPRESS AS THE FUNCTION cae Jax THAT MEANS, On, THATS wily YOUTT wars GONG BY 0 RADIANS, SHE 19 AT cloner att ONINSIDE HIS HORIZONTAL POSITION HEAD? DETERMINED BY cos 0 f You Berres REMEMBER THSL IN THE SAME Way, THE DANCER'S VERTICAL POSITION CAN Be EXPRESSED AS THE FUNCTION 120 CHAPTER 4 LET'S LEARN INTEGRATION TECHNIGUES!eA aa YES! AS 6 BECOMES LARGER, THE VALUE OF cos 0 CHANGES FROM 1, GRADUALLY BECOMES SMALLER UNTIL IT'S 0, GOES ALL THE WAY DOWN TO -1, BACK TO 0, THEN. BACK TO 1 AGAIN! BEAUTIFULI? RIGHT. AND SINCE TRIGONOMETRIC FUNCTIONS EXPRESS| WAVES, THEY CAN BE USED AS A TOOL. FOR CLARIFYING MANY THINGS IN REALLY 1 TALKING ABOUT (ails THEM, AND THEY'RE BEAMING! USING TRIGONOMETRIC FUNCTIONS 121pipyoveow Tar Vo NETH OF SHADOW OF A a THEN, LET'S FIND THIS: DRUMSTICK EQUALS z THE LENGTH OF THE / 1+ ACCURATELY. STICK MULTIPLIED By "1° JAGUELY. 122 CHAPTER 4 LETS LEARN INTEGRATION TECHNIGUES!THE SUN IS SHINING STRAIGHT DOWN ON STICK AB, WHICH IS STANDING TILTED AT ANGLE 0 FROM THE ‘GROUND. IF WE ASSUME THE RESULTING SHADOW (THE ORTHOGONAL PROJECTION) TO BE AC, THE LENGTH OF SHADOW AC EQUALS THE LENGTH OF STICK AB MULTIPLIED BY cos 0. WE CAN THINK OF THE STICK IN TERMS OF THAT'S RIGHT! ‘A FUNCTION. COSINE AND BY DEFINITION, EXPRESSES HOW MUCH SHORTER’ cosy = AC shadow) |] THe SHADOW Is ‘AB (stick) THAN THE STICK ITSELF! 20 THE SHADOWS LENGTH IS AB x cos 0. SING TRIGONOMETRIC FINCTIONS 125INCIDENTALLY, SINCE THE IN OTHER WORDS, YAXIS COINCIDES WITH ) THE Y-AXIS WHEN IT IS ROTATED BY 90 DEGREES G RADIANS), WE CAN ‘SAY sin 015 A FUNCTION THAT OUTPUTS, DELAYED By 3, THE SAME VALUES, sin| 0+ )=coso re) Sin(9+B)=Cos8 Cos(8+8) UH.WILL You eive! Us BACK OUR DRUMSTICKS? 124 CHAPTER 4 LET'S LEARN INTEGRATION TECHNIQUES!USING INTEGRALS WITH TRIGONOMETRIC FUNCTIONS: HERE ARE SPECIAL ‘SEATS FOR YOU. BE CAREFUL NOT TO. FALL, REPORTERS, AND TAKE GOOD YOUR ACTIONS ARE TOTALLY DIFFERENT FROM WHAT ‘YOU SAY. IT'S EASIER TO IN FACT, INTEGRALS ARE | UNDERSTAND IF EASIER TO OBTAIN THAN DERIVATIVES, USING INTEGRALS WITH TRIGONOMETRIC FINCTIONS 125WHAT WE NEED TO DO IS TO FIND OUT WHAT LOOKING AT THIS: cos @ x AP = COS 8 (8; ~ Op) + COS (8) ~ 0) + PUTS ME IN A +008 0), (0,~ 0.1) BECOMES. FOS. LOOK AT THIS FIGURE. DOESN'T Acangle THIS GIVE YOU A cn the ants (2000 IDEA? THIS = ‘SHOWS THAT THE INTERSECTING ANGLE OF THE Y-AXIS WITH THE TANGENT LINE PQ, WHERE P 15 THE POINT MOVED FROM (1, 0) BY ANGLE 0, 15 ALSO 6 ‘The change in cos #is the length AA, ‘That length is the orthogonal projection A,4,- Length A, ~ are A,A, x c08 0, = (0, ~ 4) x 008 8, FUTOSHI WHY DOES HE GET To BAT CHOW MEIN WHILE T HAVE TO LEARN ABOUT INTEGRALS? 126 CHAPTER 4 LETS LEARN INTEGRATION TECHNIGUES!A, (€08 0,, sin 0,) = (cos sin a) Cm. Lev’s USE THIS TO INTEGRATE FROM 070 a. A, (cos 4, sin 0) A, (c08 #,, sin 0) = (1, 0) 008 0\9 when is changed from 0 to « RIGHT! IF WE cov i= 0) v2006,0,-0) 9 9000e,,u.-e.g || ume anese | WE UND THAT NENTELY OF COSINE IS MAVA, AA, +t AY AL = AA = in So cosd = sind&— sind DERIVATIVE OF SINE IS COSINE? REMEMBER THESE FORMULAS. USING INTEGRALS WITH TRIGONOMETRIC FUNCTIONS 127FORMULA 4-4: THE DIFFERENTIATION AND INTEGRATION OF TRIGONOMETRIC FUNCTIONS Since © {" cosdd0 = sin -sin0, we know that sine must be cosine’s derivative. © (sindy =cosé = (feat 2) Now, substitute 0+ 5 for Yin ©. We get |sin{ 0+ ] = con{0 0% Using the equations from page 124, Lo a) wwe then know that © (cos#) =-sing We find that differentiating or integrating sine gives cosine and vice versa. ALL RIGHT! LET'S. DO THE CALCULUS DANCE SONGI! ALC. A CALCULUS DANCE SONG TRIGONOMETRIC VERSION 128 CHAPTER A LET'S LEARN INTEGZATION TECHNIGUES!a wee THE DANCE SONG MAKES BORING LOGIC EASIER! CALC, CALCULUS. YAY! VSING INTEGRALS WITH TRIGONOMETRIC FINCTIONS 12YEAH, WELL, YOU'RE THE ONE WEARING DANCING CLOTHES! 130. CHAPTER 4 LET'S LBARN IKTEGRATION TECHNIQUES!USING EXPONENTIAL AND LOGARITHMIC FUNCTIONS : eee 4 Fk FCs AND THE INTERNET HAVE REALLY CHANGED REPORTERS’ BY THE WAY... USING EXPONENTIAL AND LOGARITHMIC FUNCTIONS. 151SINCE COMPUTERS HANDLE INFORMATION IN THE BINARY SYSTEM, ONE BIT CAN REPRESENT TWO NUMBERS CO AND 1; TWO BITS CAN REPRESENT FOUR (00, 01, 10, AND 11); THREE BITS CAN REPRESENT EIGHT; AND 1 BITS CORRESPOND TO 2" POSSIBLE NUMBERS. EXPONENTIAL FUNCTION? AN EXPONENTIAL FUNCTION CAN EXPRESS AN. INCREASE LIKE ECONOMIC IN THE 19505 IN JAPAN, WE HAD A HIGH RATE OF ECONOMIC. ‘GROWTH: ABOUT 10 PERCENT A YEAR, 152 CHAPTER 4 LET'S LEARN INTEGRATION TECHNIQUES! [ 2E we suprose sy IS THE NUMBER OF VALUES THAT CAN BE EXPRESSED BY x BITS, THEN fix) = 2°, WHICH 15 AN EXPONENTIAL FUNCTION. EXPONENTIAL LET Me SEE... FOR EXAMPLE... A PERSON WITH AN ANNUAL INCOME OF ¥5 MILLION ONE YEAR EARNED ¥5.5 MILLION ‘THE NEXT YEAR. HIS SALARY INCREASED 10 PERCENT, AND HE COULD ENJOY 10 PERCENT MORE COMMODITIES AND ‘SERVICES THAN IN THE PREVIOUS YEAR.DON'T GET T00 EXCITED. WE HAD SUCH GOOD Days! I WOULD HAVE BOUGHT A WHOLE NEW WARDROBE AND LOTS OF OTHER THINGS! ‘SUPPOSE THE ECONOMIC ‘GROWTH I5 10 PERCENT, AND THE PRESENT GROSS DOMESTIC PRODUCT 15 Go. IN A FEW YEARS, FP WILL CHANGE AS FOLLOWS. THEN, WHAT 1S THE GROSS DOMESTIC PRODUCT AFTER 1m YEARS IN G,=G,x11 Gross domestic product after 1 year G,=G, x11 =G, x 11" Gross domestic product after 2 years Gy = Gy x LY, OR 145 Gy = Gx 11° TUES Gy. SO'THE GDP Herprodiiels lac 8 peacs NEARLY DOUBLED IN Gross domestic product after 3 y pay DOLE ed G,=G,) x11 Gross domestic product after 4 years G,=G, x 11° Gross domestic product after 5 years 0, A FUNCTION IN AN ECONOMY HAVING ‘A FORM LIKE AN ANNUAL GROWTH fix) = aoa* RATE OF a IS 1S CALLED AN EXPRESSED WITH THE EXPONENTIAL EXPONENTIAL FUNCTION FUNCTION. Hix) = al + af. USING EXPONENTIAL AND LOGARITHMIC FUNCTIONS 133BITS ARE Al ae FUNCTION CALI BACK INTO IS EASY-YoU Just |] 2PATTERNS 18 NEED TO THINK THE OTHER WAY AROUND. / 4 perreans 2075 PATTERNS =» 3.8m5 NOW, ASSUME gi) 15 THE INVERSE FUNCTION OF fix), WHICH TURNS y PATTERNS BACK INTO BITS. TRY TT. REMEMBER NOW THAT THE INVERSE theron Sea | log EXPRESSED AS \_gly) = logay: THE SYMBOL log. 124 CHAPTER 4 LET'S LEARN INTEGRATION TECHNIGUES! AN EXPONENTIAL FUNCTION. IF x BITS FUNCTION 1 yOu THA Bits CORRESPOND TO fix} ARE CODES FOR POSSIBLE NUMBERS, EXPRESSING THEN fix) = 2°. YOU KNOW, THERE Is A INVERSE FUNCTION, WHICH TURNS WHAT YOU CALLED PATTERNS, EXPONENTIAL FUNCTION Is CALLED | Acooaeric | aN FINCTION AND ‘a mae aBove \ EXPRESSED WITH eo) INVERSE LED AN ers. 90, WE CAN REPRESENT 2” POSSIBLE NUMBERS USING 20, THE RELATIONSHIP BETWEEN S AND g CAN BE EXPRESSED AS gl fx) = x AND figty)) = y- RIGHT, AND tog,2 1og,4 = 2, log,8 = 3,GENERALIZING EXPONENTIAL AND LOGARITHMIC FUNCTIONS ALTHOUGH EXPONENTIAL AND LOGARITHMIC FUNCTIONS ARE CONVENIENT, OUR DEFINITION OF THEM UP TO NOW ALLOWS ONLY NATURAL NUMBERS: FOR x IN fix) = 2° AND THE POWERS OF 2 FOR y IN gly) = log, y. WE DON'T HAVE A DEFINTION THE -8th POWER, THE “/grd POWER OR THE 2th POWER, log,5, OR log,r. TWILL TELL YOU HOW WE DEFINE EXPONENTIAL AND LOGARITHMIC FUNCTIONS IN HMM, WHAT DO AQ GENERAL, USING EXAMPLES. WE DO, THEN He GLAD THAT YOU ASKED AM I. ‘THE POWER OF CALCULUS WE USE FIRST, USING OUR EARLIER EXAMPLE, LET'S CHANGE THE ECONOMY'S ‘AKNUAL GROWTH RATE TO ITs INSTANTANEOUS GROWTH RATE. Present value (x) THIS 1S THE EXPRESSION We START WITH. USING EXPONENTIAL AND LOGARITHMIC FUNCTIONS. 135 Annual growth rate =NOW WE DEVELOP THIS INTO THE INSTANTANEOUS GROWTH RATE, AS FOLLOWS. Instantaneous growth rate Value slightly later ~ Present value | 1, Present value = ldeatization of { (Let e)-Slx) a | = Result obtained by letting ¢ +0 in 1_( S(x+e)- F(x) cml rt I(x) 50, WE DEFINE THE INSTANTANEOUS: Now, let's consider a function that satisfies the instantaneous growth rate when it is constant, or =c where cis a constant. Here we assume c = 1, and we will find fix) that satisfies Sx) I(x) 1. We first guess this is an exponential function. SINCE J"(x)= F(x) © J'(0)= 5(0) NOW, RECALL THAT WHEN hi WAS CLOSE ENOUGH TO ZERO, We HAD f(n) = #’(0)(n-0)+ F(0) 136 CHAPTER 4 LET'S LEARN INTEGRATION TECHNIQUES!From ®, we have f(h)= f(0)h+ (0) and get © s(h)=F(0)(n+1) Ix is close enough to h, we have S(x)=F(n)(x—h) + F(h) Replacing x with 2h and using f(t) =(h), J (2h) = F'(h) (2h —h) + F(h) S (2h) = F(h)(h) + f(r) F(ah) = f(h)(t+1) We'll then substitute f(h)= f(0)(h +1) into our equation. F (2h) = f(0)(R-+1)(+1) J (2h) = s(0)(n-+4)° In the same way, we substitute Sh, 4h, 6, .» forxand allow mh = 1. S (A) =F (mh) = f(0)(-+1)" similarly, J(2)=F(2mh) = F(0)(n+1)" = F(0)f(a-+n)"P (8) = s(amn)~ $(0)(n+2)" = s(0}{(1+ny"} Thus, we get S(n)=f(0)a" where we used a= (1 +h)" which is suggestive of an exponential function.” ae + Since mh = 1, Itwe tet m-yonere, {1+ 4)" +, or Euler's = Fe Then, J(1)=s(0)[ 1+ ‘number, a number about equal to 2.718. Thus, 1) =J[0) «e, which is consistent with the dis- ‘cussion on page 141 USING EXPONENTIAL AND LOGARITHMIC FUNCTIONS 1572 Next we will ind out that fix) surely exists and what itis like. EXPRESS THE INVERSE FUNCTION FROM J'(x) = fix) INDICATED ON PAGE 136, THE DERIVATIVE (OF Jx) 15 ITSELF. BUT THIS DOES NOT HELP US. WHAT IS THE DERIVATIVE OF gly) THEN? +Since we get this generally,” ‘we get this result, which ‘shows that the derivative of *~ the inverse function oy) is Now, wecan use the Fundamenta: aga ‘Theorem of Calculus. Iegives [eed since we now know gtu)=2, | i [function te) is found to ¥” | 2 (a)-9(2) + }he a function obtained by “integrating } from 1 to a. y © As shown on page 75, ithe inverse function of y =x) is x= gly) Sts) o'(y 138 CHAPTER A LET'S LEARN INTEGRATION TECHNIGUES!THIS IS A GRAPH OF INVERSE PROPORTION. LET'S DEFINE g(a) AS THE AREA BETWEEN THIS GRAPH AND THE Y-AXIS IN THE INTERVAL FROM 170 a. THIS 15 A WELL- DEFINED FUNCTION. IN OTHER WORDS, g(a) 15 STRICTLY DEFINED FOR ANY a, WHETHER IT IS A FRACTION OR v2. a1 ol ‘ Since g(1) =! su of yt = g(a)~g(1) which satisfies @. Thus, we have found out the inverse function g(y), the area under the curve, which also gives the original function flv. PLEASE TELL \ ; THE RECENT L Youre GROWTH RATE |S) 1 WONT BE CRYING! 35 IT OF THE ASAGAKE ‘SURPRISED. SING EXPONENTIAL AND LOGARITHMIC FUNCTIONS. 139SUMMARY OF EXPONENTIAL AND LOGARITHMIC FUNCTIONS Fa) 8 thowght to be the growth rts © =y=f{x) which satisfies L(x) is the function that has a constant growth fate of 1. St) ‘This is an exponential function and satisfies S'(x)= F(x) © Ifthe inverse function of y = f(x) is given by x= g{y), we have Area =1 — 1 ——e We define e (the base of the natural logarithm) as y that satisfies gly) = 1. That is, it is the « for which the area eis an irrational number spas ice between the 1/y curve and Se ee the y-axis in the interval from 1 to « equals 1. Vo CHAPTER 4 LET'S LEAAN INTEGRATION TECHNIGUES!Since f{x) is an exponential function, we can write, using constant do, F(x) Since flg(1)) =/(0) = aya’ = a, and flg(t)) = 1, we get S(g())=1=4 ‘And so we know S(e) Similarly, since J(9(e))=s(1)=a' ana S(g(e))=€ e=-a' Thus, we have f(x) =e". ‘The inverse function gly) of this is log. y, which can be simply written as Iny (In stands for the natural logarithm). Now let's rewrite © through 0 in terms of e* and In y. © To define 2*, a function of bits, for any real number x, we look at J(x)=e" (eis any real number) ‘The reason is as follows. Because * and In y are inverse functions to each other, ‘Therefore, for any natural number x, we have Jtx)=(e" ‘SUMMARY OF EXPONENTIAL AND LOGARITHMIC FUNCTIONS. 4ne MORE APPLICATIONS OF THE FUNDAMENTAL THEOREM Other functions can be expressed in the form of flx) = x", Some of them are For such functions in general, the formula we found earlier holds true. FORMULA 4-2: THE POWER RULE OF DIFFERENTIATION S(xj=x"S'(x)=ax"* PROOF: Let's express fix] in terms of e. Noting e'"* = x, we have lx em ze Thus, In f(x)=ainx Differentiating both sides, remembering that the derivative of In w and applying the chain rule, 2 CHAPTER 4 LET'S LEARN INTEGRATION TECHNIGUES!INTEGRATION BY PARTS If hx) = fix) glx), we get from the product rule of differentiation, (x)= S'(x) 9(x) + S(x)9'(x) Thus, since the function (the antiderivative) that gives f(x) g(x) +f gx) after differentiation is fx) g(x), we obtain from the Fundamental Theorem of Calculus, [LAF (e)9(x) + F(x) 9'(x)} ax = F(d)9(b)- F(a) g(a) Using the sum rule of integration, we obtain the following formula. FORMULA 4-3: INTEGRATION BY PARTS LF (x)a(x)ae + f? F(x) 9" (x) ax =F (b)9(b)- F(a) (a) As an example, let's calculate: [fxsinxax ‘We guess the integral's answer will be a similar form to x cos x, so we say fix) = x and g(x) = cos x. So we try, ff x’cosxde + J x(cosx) ax = f(x) g(x), We can evaluate that = F(x) 9(x)- (0) 9(0) ‘Substituting in our original functions of f(x) and g(x), we find that =c0s2~-0cos0=x(-1)-0 We can use this result in our first equation. Jj x cosxdx+['x(cosxy dx MORE APPLICATIONS OF THE FUNDAMENTAL THEOREM 33We then get: [feos +f sinx)dx =— Rearranging it further by pulling out the negatives, we find: [feosxax ff xstnxde =x And you can see here that we have the original integral, but now we have it in terms that we can actually solve! We solve for our original function: [fxsinxdx = [cosa +2 Remember that | cos x dx = in x, and you can see that [fxsinxae=sinx{é 47 sinz-sino+2 =0-0+ ‘There you have it. EXERCISES 1. tan xis a function defined as sin x/ cos x. Obtain the derivative of tan x. 2 Calculate oo 3. Obtain such x that makes flx) = xe“ minimum. 4 Calculate ff2xinxax A clue: Suppose f(x} = x" and g(x} in x, and use integration by parts. 144 CHAPTER 4 LETS LEARN INTEGRATION TECHNIGUES!5 LET'S LEARN ABOUT TAYLOR EXPANSIONS!IMITATING WITH POLYNOMIALS = NICE TO MEET LHAVE HEARD SO YOU. ‘MUCH ABOUT YOU, MR. SEKI.THIS 1S THE SAME DATA THAT YOU USED IN YOUR ARTICLE, TSN IT? AH, YES.. WHATS THE SOURCE OF THIS DATAP NORIKO, WHAT ARE ‘YOU DOING? YOU LOOK SusPicious. 11'S FROM BURNHAM ‘CHEMICAL. WE RECEIVED THE DOCUMENT ITSELF FROM A WHISTLE- BLOWER, WE'VE ALREADY CHECKED 115 CREDIBILITY WITH OTHER SOURCES. CANT PUBLISH MY NEW STORY Yer. BUT I WILL LEND YOU THE DATA THAT I HAVE COLLECTED 50 FAR. WB CHAPTER'S LET'S LEARN ABOUT TAYLOR EXPANSIONS!WELL, YOU VE ALOT CURIOSITY. (MR. SEKI, IZM WORRIED, BURNHAM CHEMICAL IS AN IMPORTANT. ‘SPONSOR OF THE ‘ASAGAKE TIMES. IF THEIR ILLEGAL ACT IS REVEALED, TM SURE THEY WILL STOP. ‘SUPPORTING US. L THOUGHT ABOUT THIS. v%, IMITATING WrTH POLYNOMIALS 140We IMITATED FUNCTIONS TO GET ROUGH IDEAS BY SIMPLIFYING THINGS, DIDNT we? DIFFERENTIATION WAS NOTHING BUT MAKING AN IMITATING LINEAR FUNCTION. IF WE SET p = f(a) AND 4 ={Jla) FOR FUNCTION f(x), FOR EXAMPLE, WE COULD IMITATE Fis) WITH A LINEAR FUNCTION AS ‘Feed = q + ple a) VERY NEAR YES, AN EXAMPLE 1S THE CASE OF JOHNNY FANTASTIC, WHO BEGAN TO GAIN WEIGHT AGAIN BECAUSE OF HIS BREAKUP, BUT, IN OTHER CASES, WE IMITATED A FUNCTION WITH A QUADRATIC OR A ‘CUBIC FUNCTION. L HAVEN'T DONE IF YOU PAY BACK THE MONEY ‘THIS RECENTLY. AFTER 1 YEAR, YOU PAY MUL + x). 50, HERE'S IF YOU PAY BACK THE MONEY ‘ANOTHER AFTER 2 YEARS, YOU PAY. M (1 +x) (1+), IF IVS AFTER n YEARS, YOU PAY M (1 + x)". NOW, IF WE WANT TO "EXPANDY THAT FUNCTION...” ASSUME YOU BORROW M YEN AT AN ANNUAL INTEREST RATE OF x. (14x) =14 C24 C47 + 46, at Fin (r-1)} * THIS 1S THE FORMULA OF BINOMIAL EXPANSION, WHERE ,C, = AND,C, =" (n-1) (n=4). 150 GHAPTER S LET'S LEARN ABOUT TAYLOR EXPANSIONS!TAKING ONLY THE FIRST PART, WE CAN IMITATE (1+ x)" WITH LINEAR FUNCTION Lenx eae TF YOU USED THIS: ‘APPROXIMATION, YOU WOULD EASILY BORROW TOO MUCH MONEY AND SINK INTO DEBTOR'S PRISON. (L+x)' <1ene BUT. THIS IMITATION IS IN FACT TOO. ROUGH TO BE OF 50, We USE THE QUADRATIC FUNCTION, TO IMITATE. JU..JUST A MINUTE! THOUGHT TAYLOR EXPANSION APPLIED TO (OUR NEWSPAPER! JUST BEAR WITH ME FOR PN\ A MINUTE, WILL you? MTATING WITH POLYNOMIALS 151FORMULA 5-1: THE FORMULA OF QUADRATIC APPROXIMATION (+x) 14m A IF WE MODIFY THIS EXPRESSION ALITILE, WE GET A VERY INTERESTING LAW. pee ae L+ne+3(nx)' - pln) gre =140.7+ } 0.7? =1.945=2 | Nearly zero, so we negiect it, In short, if nx = 0.7, (1 + x)" is almost 2. This can be written as a law as follows. (1+x)" <14+ne+™ LAW OF DEBT HELL ‘When years to repay loan x interest rate = 0.7, the amount you will repay is about twice as much as you borrowed. "35 YEARS AT 2 PERCENT | DON'T TAKE ME YES. WE DON'T KNOW WHICH SIDE WILL APPEAR. BUT WE DO KNOW THE CHANCES (OF A PARTICULAR SIDE IS1INZ, FOOL. THE GRAPH ON TOP SHOWS ‘THE PROBABILITY OF GETTING HEADS WHEN 20 COINS ARE FLIPPED AT ONCE, PLOTTED WITH THE NUMBER OF HEADS: ‘ON THE HORIZONTAL AXIS AND THE PROBABILITY ON THE VERTICAL AXIS. 135 7 9 1118151719 ‘The number of heads when 20 coins are flipped at once (binomial distribution) 02 /SIEN on / aE ° vrei oOmiao Standard normal distribution Of, TT LOOKS LiKE THE LOWER GRAPH. YES, IT OVERLAPS WITH THE GRAPH (OF ANORMAL DISTRIBUTION ALMOST PERFECTLY. WHAT DOES TAYLOR EXPANSION TELL US? 167IN FACT, IF WE DEFINE g,(x) AG "THE PROBABILITY OF GETTING x HEADS WHEN 7 COINS ARE FLIPPED AT ONCE" AND ALLOW n TO APPROACH +20 FOR THE GRAPH OF 9,(x). G16 READ'AS INFINTY HE WROTE THE SAME: EQUATION BEFORE! HE DOESN'T HAVE TO USE TWO COASTERS! SEE THAT IT IS PROPORTIONAL PATI TO THE NORMAL FUNCTION ABOVE. * The distribution of such probabilities as that of getting x heads when n coins are flipped is generally called the binomial distribution. For example, let's find the probability of getting 3 heads when 5 coins are flipped. The probability of getting HHTHT (H: heads, T: tails) is since mer are Cs ways of gtng 9 Neds and? tats tis (2). Te general expression ie ,C,() We wil show tat im ia very lng, th nomial distribution is the normal distribution, 168 CHAPTER'S. LET'S LEARN ABOUT TAYLOR EXPANSIONS!USING THE BINOMIAL. DISTRIBUTION, SINCE THE GRAPH OF ‘fle) 1G SYMMETRICAL ABOUT x = 0 AND g,(x) ABOUT WE CONSIDER g,(") INSTEAD "2 OF g,(0- 1 ‘9q(¥) CAN BE WRITTEN IN THIS WAY. WE GET hy, THE SCALED FUNCTION J Waar DOES TAYLOR EXPANSION TELL US? 164= ee = Se ‘Vn | DEVIATION. IF YOU DON'T KNOW STATISTICS, SIMPLY REGARD IT AS A MAGIC WORD! WELL, WENOW = fy CONVERT THE UNIT INTO SINCE x 15 AWAY FROM THE CENTER 7, * STANDARD DEVIATION 'S AN INDEX We USE TO DESCRIBE THE SCATTERING OF DATA, 70 CHAPTERS LET'S LEARN ABOUT TAYLOR EXPANSIONS!I THis Way, We CHANGE THe VARAOLE, THE NEW ONE, 219 THE NUMBER OF STANDARD DevianIONs AWAY FROM THE CENTER, ‘THis, BUT SHALL WE MOVE ON TO es de (De WHAT DOES TAYLOR EXPANSION TELL USP 171THANK YOU. I THINK WE'RE DONE. THERE ARE ONLY A 1 GUESS 1 SHOULD BE HAPPY 1 STILL Approximating In (m!) Area =1n m. yrinx lnm!=Inl¢In21n3+..+Inm * If we pack rectangles in the graph of n x, as shown here, we get : lam Ind.» f"Inxae (xnxx) stmt xxt x mm Thus, [f'n xdx = (minm—m)-(1in1-1) =minm-m+1 Since we will use this where m is very large, m In mis the important term. -m + Lis much smaller, so we'll ignore it. Therefore, we can consider roughly that In m! = min m. 172 CHAPTERS LETS LEARN ABOUT TAYLOR EXPANSIONS!WELL, LET'S JUST FINISH QUICK, GINE THESE TO Usl! AHHHH! Dr hyde on B+ nt (9+ 2) Oo (Ba) -B 2) tn 8-F 2) AFTER A LOT OF ALGEBRA, WE GET a,c -[(B+5e) tn (0 z)+(B-Fe) th (Ba) nf {2 WE USED, E.G, in) Be NOW, LET'S USE A TAYLOR EXPANSION, WHICH YOU'VE BEEN | WAITING FOR. LHAVEN'T BEEN WAITING FoR IT. WHAT DOES TAYLOR EXPANSION TELL USP 173WHEN t IS E - close TO 1 ZERO, J] Now, -n “in IS VERY, MIn(leet-st? ft. CLOSE TO ZERO IF nis LARGE ENOUGH. cquapeanic APPROXIMATION)” THEREFORE, 74 CHAPTER S LET'S LEARN ABOUT TAYLOR EXPANSIONS!‘SINCE WE NOW KNOW Inn, (x) =-324, We GET h, (x)=-e 2", THAT'S I IF YOU ARE AFRAID THAT THE HIGHER-DEGREE TERMS OF x? AND MORE THAT APPEAR IN THE TAYLOR EXPANSION (OF In MIGHT AFFECT THE SHAPE OF h(x) (re LARGE ENOUGH), ACTUALLY CALCULATE n(x), USING feel a+e)~t-de rte Indedat-ge+y YOU WILL FIND THAT THE TERM OF 2* HAS n IN THE DENOMINATOR OF ITS COEFFICIENT AND CONVERGES TO ZERO, OR DISAPPEARS, WHEN n> on. AS FOR THE NORMAL DISTRIBUTIONS, CAN WE APPLY THEM TO THINGS OTHER THAN COIN FLIPPING? A WHET D085 TAYLOR EXPANSION TELL US? 75ARE YOU THINKING 11's out ABOUT APPLYING OUR OF THE STUDIES TO Love AGAIN? | “|| avestion PROBABILITY CAN ONLY : APPLY WHEN PHENOMENA ‘ARE UNINTENTIONAL AND, PURELY RANDOM. HOW ABOUT IN THE CASE OF UNINTENTIONAL AND LISTEN IF WE DARE TO. ASSUME VERY ROUGHLY THAT ‘THE WAY TWO PEOPLE FALL IN LOVE 1S SOMETHING LIKE THE COMBINATION OF THE RESULTS (OF FLIPPING AN INFINITE NUMBER OF COINS. WELL, SINCE WE HAVE FOUND THAT THE DISTRIBUTION OF THE RESULTS OF COIN FLIPPING 15 APPROXIMATELY A NORMAL DISTRIBUTION, TT WOULD NOT BE ‘SURPRISING IF A NORMAL DISTRIBUTION COULD BE CALCULATED FOR LOVE. 6 CHAPTERS. LET'S LEARN ABOUT TAYLOR EXPANSIONS!PROBABILITY 16 LIMITED ONLY TO UNCERTAIN PHENOMENA THAT ALLOW NO INTENTIONALITY. IX) IM SORRY FOR BEING oriNeRe Ie AVERY Wty D0 You Neve mice ome, UNDERSTAND WEP WHAT DOES TAYLOR EXPANSION TELL USP 177EXERCISES 1. Obtain the Taylor expansion of fix) = €™ at x = 0. 2. Obtain the quadratic approximation of f(x) atx= cosx 3. Derive for yourself the formula for the Taylor expansion of f{x) centered at. x= 1, which is given on page 159. In other words, work out what c, must be in the equation: F(x) = cy +¢,(x-a) +6, (x-a)’ +..+¢,(x-a)" 7S CHAPTER'S LB1'S LBARN ABOUT TAYLOR EXPANSIONS!LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!WHAT ARE MULTIVARIABLE FUNCTIONS? MR. SEKI IS GOING BACK TO THE MAINZ OFFICE?CAUSE AND EFFECT... REMEMBER THAT. WE TALKED ABOUT THAT IN ONE OF OUR FIRST LESSONS, BUT YOU TOLD ME, “AN EFFECT OCCURS BECAUSE IT HAS A EVERY DAY! T EVEN HAD NIGHTMARES: RELATIONSHIP CAN BUT THIS TRANSFER HAS REMINDED ME THAT THE WORLD IS NOT SO b/ 31'S TRUE THAT WE HAVE BEEN EXPLORING SIMPLE FUNCTIONS THAT HAVE A CAUSE AND AN EFFECT. I GUESS MY TRANSFER TO THE MAIN OFFICE HAS BEEN BROUGHT ABOUT AS ‘A COMBINED RESULT OF ‘SEVERAL CAUSES, WHAT ARE MULTVARWBLE FUNCTIONS? 161TT CAN BE EXPRESSED ‘BY THESE DIAGRAMS. IN THE CASE OF NORIKO, x, IN THE CASE OF fez, SER 2 6 LAST MONTH'S BLUNDER, x, 15 THIS MONTH'S BLUNDER, "AND x, AND x, ARE POOR EXCELLENT WRITING, ¥y 16 HARD-HITTING GROOMING AND HYGIENE, WHICH MAKES y HER REPORTING, AND 2.15 TRANSFER TO. WELL, 1 DONT HE MAIN OFFICE. 15, KNOW THE DEMOTION TO WRITING, REASONS FOR OBTUARIES. oe MY TRANSFER ALL RIGHT, THAT'S ENOUGH. NORIKO, WE DON'T HAVE MUCH TIME LEFT. LET'S LEARN THE BASICS GUICKLY. 182 GUAPTER 6 LET'S LEARN ABOUT PRETIAL DIFFERENTIATION!THE FUNCTION OF IWILL Give you ip WRITEN AS Some EXAMPLES OF 2= glx, y), AND FUNCTIONS THAT HAVE THAT OF THE RIGHT DIAGRAM IS WRITTEN TWO CAUSES, THAT AS y= MX, Xp ye Ky) 15, TWO-VARIABLE FUNCTIONS. EXAMPLE t Assume that an object is at height h(v, é) in meters after t seconds when itis thrown vertically upward from the ground with velocity v. Then, h(v, t) is given by h(v.t)=0t-4.9¢° exampte z The concentration fix, y) of sugar syrup obtained by dissolving y grams of sugar in x grams of water is given by F(y)= 100 x+y EXAMPLE 3 When the amount of equipment and machinery (called capital) in a nation is represented with K and the amount of labor by L, we assume that the total production of commodities (GDP: Gross Domestic Product) is given by YU, K). IN ECONOMICS, ¥ (L,I) = BLK (CALLED THE COBB-DOUGLAS FUNCTION) QNHERE a AND i ARE CONSTANTS) IS USED AS AN APPROXIMATE FUNCTION OF VIL, K). SEE PAGE 203. examrts 4 In physics, when the pressure of an ideal gas is given by P and its volume by ¥V its temperature T is known to be a function of Pand Vas T(P, V). And it is given by T(P.V) WHAT ARE MULTWARIABLE FUNCTIONS? 163THE BASICS OF VARIABLE LINEAR FUNCTIONS WHAT DO YOU THINK THE PROPERTIES OF THESE COMPLICATED TWO-VARIABLE FUNCTIONS? TWO-VARIABLE LINEAR FUNCTIONS ARE GIVEN IN A FORM LIKE 2 = lx, y) = ax + Dy +0 (WHERE a, b, AND c, ARE CONSTANTS.) NOW, LET'S SEE WHAT THEIR GRAPHS LOOK LIKE. SINCE THEY HAVE TWO INPUTS (x AND w) AND AN OUTPUT (2), IT 15 NATURAL TO USE 3-DIMENSIONAL ‘COORDINATES. WE D0 70 EXAMINE DO We USE IMITATING LINEAR WELL, YES. BUT SINCE WE NOW HAVE TWO-VARIABLE FUNCTIONS, WE HAVE TO USE TWO-VARIABLE LINEAR FUNCTIONS. WELL, JUST THINK (OF AN IMAGE IN WHICH THE XY PLANE |S THE FLOOR AND THE Z-AXIS IS A PILLAR. 184 CHAPTER 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!IS SOMETHING WRONG? YOU SEE, THE POINT P AT THE COORDINATES: (2, 8, 5) 1S THE PONT AT THE TOP OF A STICK STANDING AT (2, 3) ON THE FLOOR AND HAVING A LENGTH OH! NO, NOTHING. LET'S CONTINUE. FIRST, WE PLACE A STICK HAVING THE LENGTH f(1, 2)=3x14+2%2+1=8 AT POINT (1, 2) ON THE FLOOR. JIN THE SAME WAY, THE HEIGHT. OF THE GRAPH HAS A VALUE OF S43) =3%442%341=19 AT POINT (4, 3) Now, WHAT DO YOU. THINK THE GRAPH OF THE TWO-VARIABLE LINEAR FUNCTION 2=flx, y)=ax+by +e LOOKS LiKE?” Lev’s Draw ‘THE GRAPH OF 2=flx, y) =x +2y 41 ‘AS 'AN EXAMPLE. + ALTHOUGH WE SHOULD ACTUALLY WRITE ITA (4, 3,0}, WELL USE (4, 8) FOR SIMPLICITY. THE BASICS OF VARIABLE LINEAR FUNCTIONS 185IN THE SAME WAY, WE PUT UP 16 STICKS AT 16 POINTS LOOKING AT THIS. be, y) SATISFYING 1 0 when ¢ +0” means the following: mm J+ eb) J (ab) This is the slope of DE. Here, we should realize that the left side of this expression is the same as single-variable differentiation. In other words, if we substitute b for y and keep it constant, we obtain f(x, b), which is a function of x only. The left side of @ is then the calculation of finding the derivative of this function at x = a. Although we are very much tempted to write the left side as f(a, b) since itis a derivative, it would then be impossible to tell with respect to which, x or y, we differentiated it So, we write “the derivative of f obtained at x = a while y is fixed at b” as f(a, b). Jeers, J, 18 called “the partial derivative of fin the direction of x". This is the notation corresponding to the “prime” in single-variable differentiation. a The notation gy (a, b), that corresponds to 5x, is also used. In short, we have the following: “The derivative of fin the direction of x obtained at x = a while y is fixed at b" J.(ab)=Z (a,b) also writen as | [ = Slope of DE In exactly the same way, we can obtain the following. “The derivative of fin the direction of y obtained at y = b while x is fixed at a” PARTIAL DIFFERENTIATION 103,We have now found the following. If z= fix, y) has an imitating linear function near (x, y) given by © z= J, (a,b)(x-a)+ J, (a.b)(y-b)+ f(a.b) = (a, b), itis on x Ean)(x-a)+H a.r)(y-b)+J b) Consider a point («, ) on a circle with radius 1 centered at the origin of the x- y plane (the floor). We have a? + f° = 1 (or a= cos @ and B= sin @). We now calculate the derivative in the direction from (0, 0) to (a, f). A displacement of distance ¢ in this direction is expressed as (a, b) (a+ at, b+ fit). Ifwe set c= at and 6 in ©, we get relative ror = L(@*atb+ Bt) - F(ab) (pat-+aBt) Vee + pre Ee ae Setar) SOF) os © since Ja? +f?=1 Assuming p = J,(a, b) and q =J,(a, b), we modify @ as follows: @ Slaratbdpt)—sa.b+ pt), $(ab+ pt)-F(a.b) . ; ~S.l@sb)a~ J, (asb)p Since the derivative of fix, b+ ft) a function of x only, at x= ais J. (a,b + pt) we get, from the imitating single-variable linear function, F(a+at,b+ pe)- f(a,b+ Bt) = J, (ab+ pt)at We nave calculated the imitating linear function tn such a way that ite relative error fpprosches O when AP» in theory direction, It ie not apparent, however, ifthe relative rror~>O when AP > 0m any airetion for the linear function that is made up ofthe deriva- tives f(a, b) and f(a, b). Well now look into this in detail, although the discussion here will snot be so strict 194 CHAPTER 6 LETS LEARN ABOUT PARTIAL DIFFERENTIATION!Similarly, for y we get S(a.b+ Bt) - F(a,b)= f, (a,b) Be Substituting this in ©. © = 4. (a.b+pt)a+ J, (ab) pt-f,(ab)a~J,(a,b)p =(4.(ab+ ft)-f, (a.b))a Since J,(a, b + [it) -f,{a, b) = 0 if tis close enough to 0, the relative error = @ = 0. Thus, we have shown “the relative error > 0 when AP > 0 in any direction.” It should be noted that f, must be continuous to say f,(a, b + fit) - f,(a, b) = 0 (t ~0). Unless it is continuous, we don’t know whether the derivative exists in every direction, even though f, and f, exist. Since such functions are rather exceptional, however, we won't cover them in this book. EXAMPLES (FUNCTION OF EXAMPLE | FROM PAGE 185) Let's find the partial derivatives of h(v, t) = vt ~ 4.9¢° at (0, ) = (100, 5). In the v direction, we differentiate h(o, 5) = Sv ~ 122.5 and get oh v,5)=5 Thus, an, ae (100.5) =n, (100.5) =5 In the ¢ direction, we differentiate h(100, ) = 100t - 4.91? and get ah 28 (400,t) =100- Gp (1004t) = 100 - 8.8 ah i 57 (200.5) =f, (100,5) = 1009.85 =51 And the imitating linear function is L (x.y) =5(v-100) + 51(t~5)~377.5 PARTIAL DIFFERENTIATION 195In general, ah _, ah av - 9.8 ‘Therefore, from @ on page 194, near (v, t) = (Uo. fal- Ia(vit) = ty (v- 04) + (Up ~ 9.8ty) (t= fy) +R (Yost) Next, we'll try imitating the concentration of sugar syrup given y grams of sugar in x grams of water. oy x 100(x+y)-100yx1 __100x ay (x+y) Thus, near (x, y) = (a, b), we have DEFINITION OF PARTIAL DIFFERENTIATION When z =/lx, y) is partially differentiable with respect to x for every point (, y) ima region, the function (x, y) fx, yh, which relates (x, y) to f,lx. Ys the partial derivative at that point with respect to x, is called the partial dif- ferential function of z = fix, y) with respect to x and can be expressed by any of the following: So S(%Y) Similarly, when z = flx, y) is partially differentiable with respect to y for every point (x, y) in the region, the function (xy) > F, (xy) 196 GUAPTER 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!is called the partial differential function of z =Jtx, y) with respect to y and is expressed by any of the following: of os Se LiOW Sy ay Obtaining the partial derivatives ofa function is called partially differentiating it TOTAL DIFFERENTIALS TOTAL DIFFERENTIALS From the imitating linear function of z = flx, y) at (x; y) = (a, b), we have found S(%u) =F, (asb)(x-a)+ J, (arb)(x—b)+ (a,b) ‘We now modify this as @ $(x.y)~S(a.b)~ Ta.v)(x-a)+Fla.v)(y-») Since fix, y) - f(a, b) means the inere- ment of 2 = fis, y) when a point moves from (a, b) to (x, y), we write this as Az, as we did for the single-variable functions. Also, (x a) is Ax and (y - b) is Ay. ‘Then, expression @ can be written as 25 2 dan ac 2a a ane ae cau ae ‘This expression means, “If x increases from a by Ax and y from b by Ay in z =x, y}, z increases by ae a seo pou TOTAL DIFFERENTIALS 197since @ Axis “the increment of z in the x direction when y is fixed at b” oz Ok and 5; 4y is “the inerement in the y direction when x is fixed at a,” expres- sion @ also means “the increment of z = flx, y) is the sum of the increment in the x direction and that in the y direction.” When expression @ is idealized (made instantaneous), we have a2 a 2 © a= Fay ay EXPRESSION © OR © IS CALLED THE FORMULA OF or THE TOTAL DIFFERENTIAL. © df= S.de+ Say (has been changed to d.) ‘The meaning of the formula is as follows. Increment of height of a curved surface Partial derivative , Increment in, Partial derivative , Increment in in the x direction * the x direction * in the y direction “ the y direction Now, let's look at the expression of a total differential from Example 4 (page 183). By converting the unit properly, we rewrite the equation of temperature as T= PV. ar _a(pv)_ z Vv and ap oP ‘Thus, the total differential can be written as dT = VAP + Pav. In the form of an approximate expression, this is AT ~ VAP + PAV. THIS MEANS THAT FOR AN IDEAL GAS, THE INCREMENT OF TEMPERATURE CAN B= CALCULATED BY THE VOLUME TIMES THE INCREMENT OF PRESSURE PLUS THE PRESSURE TIMES THE INCREMENT OF VOLUME. E Teconstant Volume V 198 CHAPTER 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!CONDITIONS FOR EXTREMA WHAT A VIEW! SANDA HASN'T CHANGED AT ALL! (Ol, YOU STARTED THE LESSON ALREADY? IF WE LOOK AT THAT MOUNTAIN AS A TWO- VARIABLE FUNCTION, ITS TOP IS A MAXIMUM, CONDTIONS FOR EXTREMA 199‘The extrema of a two-variable function fix, y) are where its graph is at the top of a mountain or the bottom of a valley. Maximum P y y 3 minimum ° i ° * Maximum point P / — o ¥ Since the plane tangent to the graph at point P or Q is parallel to the xy plane, we should have S (xy) = p(x-a)+a(y~b)+ f(a.b) with p = q = 0 in the imitating linear function. Since b)= J, (a.b)=0 ‘The opposite of tnis is not true. In other words, even if/(a, 6)=J, (a,b) have an extremum at (x,y) = (a,b). Thus, this condition only picks'up the c 200 CHAPTER 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!AT THE EXTREMA OF A TWO- VARIABLE FUNCTION, THE PARTIAL DERIVATIVES IN BOTH THE x AND y DIRECTIONS ARE ZERO. Let's find the minimum of fix, y) algebraically. Since (x - y)* + (y - 2)°. First, we'll find it (x-yl 20 (y-2)'20 S(sy)=(e-u) +(y-2) 20 Ift'we substitute x = y= 2 here, F(22)=(2-2)" +(2-2)'=0 From this, fix, y) 2/12, 2) for all (x, y). In other words, flx, y) has a minimum of zero at (x. y) = (2. 2). On the other hand, Itwe set a ate-pyana Z-a(e-y)(-apeaiy-a)=-ae aya yy ax ay and solve these simultaneous equations, 2x-2y=0 -2x+4y-4=0) wwe find that (x, y) = (2, 2), just as we found above. THE SOLUTIONS ARE THE SAME! CONDITIONS FOR EXTREMA 201APPLYING PARTIAL DIFFERENTIATION TO ECONOMICS HE WAS A FORMER ECONOMIST, AND IN 1927, HE THOUGHT ABOUT THE PROBLEM (OF SHARING NATIONAL INCOME IN CAPITAL ‘AND LABOR. THERE WAS A SENATOR FROM ILLINOIS NAMED PAUL DOUGLAS WHO SERVED FROM $949 To 1966. THERE ARE ROUGHLY TWO TYPES OF ROUTES IN WHICH GROSS DOMESTIC PRODUCT (GDP), WHICH IS THE | AMOUNT OF PRODUCTION WITHIN A COUNTRY IN ONE YEAR, I5 SHARED ‘AMONG THE PEOPLE ‘OF THE COUNTRY, THE SECOND IS THE WAY THE FIRST ONE IS THE IN WHICH GDP Is SHARED WAY IN WHICH GDP IS AS STOCK DIVIDENDS TO SHARED AS WAGES THE OWNERS OF CAPITAL, SUCH AS MACHINERY AND EQUIPMENT. 202 CHAPTER € LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!DOUGLAS STUDIED THE ssi T'S STRANGE THAT THE TABOR AND CAPITAL RATIO WAS CONSTANT, SHARES IN THE UNTED IE bn EVEN THOUGH THE ‘STATES AND FOUND THAT ECONOMIC SITUATION ‘THEIR RATIO HAD BEEN WAS CHANGING EVERY ALMOST CONSTANT FOR MINUTE. ABOUT 40 YEARS. ABOUT 70 PERCENT (07) OF GDP WAS SHARED | AS WAGES FOR LABOR, { AND 30 PERCENT 3) AG STOCK DINIDENDS TO CAPITAL OWNERS. YOU WANT TO KNOW WHAT THE PRODUCTION FUNCTION JL, K) THAT BRINGS THIS RESULT LOOKS LiKE, DONT YOU? __, DOUGLAS WAS ‘TROUBLED T00, SO HE ASKED CHARLES COBB, A MATHEMATICIAN, ABOUT IT. THE FUNCTION THEY FOUND IS THE FAMOUS. (COBB-DOUELAS FUNCTION. BELOW, L REPRESENTS LABOR, K REPRESENTS CAPTIAL, ‘AND i AND a ARE ‘CONSTANTS. ‘AH, WILL YOU TELL ME IN MORE DETAIL ABOUT MY (COBB-DOUGLAS FUNCTION JS (LK) = BLK OKAY. THIS 15 A GOOD APPLICATION OF TWO- ‘VARIABLE FUNCTIONS. APPLYING PARTIAL DIFFERENTIATION TO ECONOMICS 203First, let's suppose that wages are measured in units of w, and capital is measured in units of r. We'll consider the production of the entire coun- try to be given by the function iL, K) and assume the country is acting as a profit-maximizing business. The profit P is given by the equation: P= S(L,K)-wh rk Because we know that a business chooses values of L and K to maximize profit (P), we get the following condition for extrema: The relations far to the right mean the following. Wages = Partial derivative of the production function with respect to L Capital share = Partial derivative of the production function with respect to K Now, the reward the people of the country receive for labor is Wage x Labor = wh. When this is 70 percent of GDP, we have © wl 75 (LK) Similarly, the reward the capital owners receive is © 1K =0.35(L,K) From © and ©, F ype @ Fx1-075(uK) From @ and ®, a @ LyK-oss(L, J xK=03(L.K) 1204 CHAPTER 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!Cobb found fiL, K) that satisfies these equations. It is S(LK) = BLK where f is a positive parameter meaning the level of technology. Let's check if this satisfies the above conditions. H , 2(OLK) coe Cun 7 LK? 7 § (Lak) xL=0.78U'K™ xt Boe OR) 2 0.5pt7K aK 1K 1.3 BLK .3.f (LK) 02) eR ‘SO, PARTIAL DIFFERENTIATION rEVEALED A MYSTERIOUS. LAW HIDING IN A LARGE- SCALE ECONOMY—RULES ‘THAT DETERMINE A COUNTRY’S WEALTH, PARTIAL DIFFERENTIATION 15 ALIVE AND WELL BEHIND THE SCENES, INT IT YES, IT SURELY APPLYING PARTAL DIFFERENTIATION TO ECONOMICS 208THE CHAIN RULE We have seen single-variable composite functions before (page 14). y= F(x), 2=9(y),2-9( F(x), a(F(x)) = 9 (Fe) (8) HERE, LET'S DERIVE THe FORMULA OF PARTIAL DIFFERENTIATION CTHE CHAIN RULE) FOR MULTARIABLE COMPOSITE FUNCTIONS. We assume that z is a two-variable function of x and y, expressed as z= Sle, y), and that x and y are both single-variable functions of t, expressed as a(t) and y = b(t), respectively. Then, z can be expressed as a function of ¢ only, as shown below. i LS» ig oes ae Ca ‘This relationship can be written as 2=S(xy)=S(a(t).b(¢)) dz What is the form of gj then? We assume al{,) = x,y bit) = Yo ANd {Ux Yo) = Slat). blt)) = Zo When t and consider only the vicinities of fy, X. Yo. and Zy. If we obtain an « that satisfies © 2-2,-ax(t-t) dz itis @ (tg). 206 CHAPTER 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!First, from the approximation of x= a(t ao © xx = (4 )(t-t) Similarly, from the approximation of y = b(0, a © yu MC) t-t Yn vo Plb)(e-b) Next, from the formula of total differential for a two-variable function Sle. y). 28, = TE (xu) 8) +E so) ) Substituting @ and © in ©, x.y) +L (erm) 2G )(E- BF 5 (Hor Mo) Ge (fo) (Ef) + 54 (Horo) gp (Fo) (t= fo) (x day), ag (0 Yo) Ge (Fo) * 5g 0) F(t) (t-%) Comparing © and ©, we get Ess) Geto) + (suv) Gell) This is what we wanted, and we now have the following formula! FORMULA 6-1: THE CHAIN RULE When z= f(x,y),x=a(t).y=D(t) dz_¥ da, ¥ db dt ax dt * ay at ‘THe CHAN RULE 207MR. SEKI, WHY DONT I GIVE YOU A LESSON Now? UH, OKAY, IT'LL BE FUN TOBE A ‘STUDENT AGAIN. HERE, WE HAVE A FACTORY OKAY! LET'S USE ‘AN ENVIZONMENTAL FROM WHICH WASTE IS. RELEASED AS A RESULT OF A MULTIVARIABLE PROBLEM! PRODUCTION OF COMMODITIES. FUNCTION TO THINK THE WASTE SUBSEQUENTLY POLLUTES THE SEA, CAUSING ‘A REDUCTION IN THE LOCAL. men He D deco] Mf q ‘SUPPOSE THAT x WORKERS AcTuines oF A ousneas Have Laser epee ripe ‘ G@IVEN BY fix), THE FACTORY ALSO ONS THROUGH THE MARKET, RELEASES WASTE AS GOODS Ag 9 me CAPE 9 CALLED AN ARE MADE, WHICH AFFECTS THE HARMFUL ExreRNALES, SUCH a ee NESTE EXTERNAL TES. TR a a Ler CALL THE f A quantity oF waste b= bi fix). NOW. 208 CHAPTER 6 LET'S LEARK ABOUT PARTIAL DIFFEREKTIATION!We assume that the catch of fish can be expressed as a two-variable function gly, b) of the amount of labor y and the amount of waste b. (The catch gly, b) decreases as b increases. Thus, “2 is negative.) Since the variable x is contained in gly, b) = gly. DUJ(«)), production at the factory influences fisheries without going through the market. This is an externality. First, let's see what happens if the factory and the fishery each act (self- ishly) only for their own benefit. If the wage is w for both of them, the price of a commodity produced at the factory p and the price of a fish q, the profit for the factory is given by ) = pf (x) — wx ‘Thus, the factory wants to maximize this, and the condition for extrema is x)-w=0 pf'(x)=w Let s be such x that satisfies this condition. Thus, we have ® pi (s)-w This s is the amount of labor employed by the factory, the amount of pro- duction is fis), and the amount of waste is given by br=b(F(s)) Next, the profit P, for the fishery is given by P, =ag(y.b)—wy Since the amount of waste from the factory is given by b* = bls) ® P= qa(yb*)-wy which is practically a single-variable function of y. To maximize P,, we use only the condition about y for extrema of a two-variable function, A ag ag 8 F=gS(y.b2)-w-0 49 (y.b4) ay ay! ) ay ‘Therefore, the optimum amount of labor t to be input satisfies NORIKO 6NES A LESSON. 200‘The production at the factory and the catch in the fishery when they act freely in this model are given by jis) and g(t, b*), respectively, where s and t satisfy the following. @ pf'(s)-w © wav(s(s)).a (0 NOW, MR. SEKI, LET'S CHECK IF THIS IS THE BEST RESULT FOR THE WHOLE SOCIETY. IF WE TAKE BOTH THE FACTORY "AND THE FISHERY INTO ACCOUNT, WE ‘SHOULD MAXIMIZE THE SUM OF THE PROFIT FOR BOTH. P= pf(x)+49(y.b(S(x))) -wx-wy ‘The first partial derivative is obtained as follows aolub(0e)) ax = wh (x) +42 (d(J(x))) 0S 09) F(x) Be wrx)ea (Here, we used the chain rule.) 210 CHAPTER 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!Thus, Fe a0es(praig(un(sta))o(s(e)] site simitany, o B-0 a (uo(s(a)) ‘Thus, if the optimum amount of labor is S for the factory and T for the fishery, they satisty [p+4a22(nv(s(s))0(s(S))|5(5) 8 aZiro(s(s)) Although these equations look complicated, they are really just two- variable simultaneous equations. If we compare these equations with equations ® and ®, we find that @ and ® are different while © and ‘ are the same. Then, how do they differ? @ pxf(s)-w @ (p+¥)xs(S)=w As you see here, ¥ has appeared in the expression. Since (¥ q a (8) ts negative, p+ w is smatter than p. Since f{S) or f(s) is multiplied to the first part to give the same value w, FS) must be larger than f(s). Slope J’ is small. Slope f’is large. \ Six) Now, SINCE ‘THE GRAPH OF ses) GENERALLY LOOKS LIKE THIS, NORIKO GIVES A LESSON 211FOR THE BENEFIT OF SOCIETY, THE FACTORY SHOULD REDUCE PRODUCTION DOWN TO S FROM s, THEIR PRODUCTION IN THE CASE OF PURELY SELFISH ACTIVITIES. 30 ARE THERE ANY ‘GOOD MEANS TO MAKE THE FACTORY VOLUNTARILY REDUCE PRODUCTION FROM sTOS? teqHe GOVERNMENT FORCES THE FACTORY TO REDUCE WHiLe THE BENEFIT OF THE SOCIETY BASICALLY REACHES A MAXIMUM AT THE INTERSECTION OF THE DEMAND CURVE, WHICH EXPRESSES SELFISH ACTITIES, AND THE SUPPLY CURVE," IT DOES NOT HAPPEN IF A NEGATIVE EXTERNALITY EXISTS, SUCH AS POLLUTION, IN THis case. A GOOD MEANS OTHER THAN THAT IS: THE GOVERNMENT ae (ES THE FACTC CALLED AN TREROPORTIONTO. ENVIRONMENTAL ITS PRODUCTION. a 212 CHAPTER 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION! TO ALLEVIATE GLOBAL WARMING, A CARBON TAX, TAXATION ON THE EMISSION OF CARBON, IS ALSO BEING DISCUSSED.THEN, THE PROFIT © IN THE CASE OF SELFISH ACTNITIES BECOMES LIKE THIS. LET'S ASSUME THAT THE TAX ON A UNIT COMMODITY PRODUCED AT THE FACTORY |S -¥ THE CONDITION FOR EXTREMA THAT MAXIMIZE THIS IS... ag, vq 6 (5(S)) @ Baap (x)-wiw s(x) 06 (+9) F(x) THis 16 A POSTIVE CONSTANT. SINCE © 1S THE SAME EQUATION ‘AS 3) THE PRODUCTION AT THE FACTORY NOW MAXIMIZES THE ‘BENEFIT FOR SOCIETY. ORDINARY TAKES pre ONCOME TAX, HAVE YOU Gor Tr, CONSUMPTION TAX) TAK Is FOR gti tercalt ‘BRE FOR PUBLIC MANTAINING A HEALTHY ENVIRONMENT BY CONTROLLING THE ECONOMY. NORKO GIVES ALESSON 25THE ONLY ONE LEAVING? I-THINK TM ALMOST DONE PACKING, ALREADY TOLD HIM. ANEW ASSIGNMENT 215You wi. BE NOTIFIED SOON DIDN'T EVEN KNOW ‘OUR COMPANY HAD [AN OKINAWA OFFICE. THIS 15 A FAREWELL FOR YOU, WRITE GOOD ARTICLES WITH TH. / 216 CHAPTER 6 LETS LEARN ABOUT PARTIAL DIFFERENTIATION!Expected Awouoog pue JusuUO TAU :SeTIIgDERIVATIVES OF IMPLICIT FUNCTIONS A point (x,y) for which a two-variable function fx, y) is equal to constant ¢ describes a graph given by f(x, y) = c. When a part of the graph is viewed as a single-varlable function y = h(a}, itis called an implicit function. An implicit function h(x) satisfies fix, Rix) ~ forall x defined. We are going to obtain A(x) here. When z = lx y), the formula of total differentials is written as dz = f,dx + ‘Jay. If (x. y) moves on the graph of fe, y)= c. the value of the function fix, y) does not change, and the increment of z is O, that is, dz = 0. Then, we get Om fide + fay. Assuming f, »O and modifying this, we get a fe aes, ‘The left side of this equation is the ideal expression of the increment of y divided by the increment of x at a point on the graph. It is exactly the derivative of h(x). Thus, rae nw ( i ewer “Ses, y) = 7, where f(x, y) = x7 + y?, describes a circle of radius r centered ‘at the origin, Near a point that satisfies x* » r', we can solve flx, y) =" + y' = 7 to find the implicit function y = his) = r - x" or y=h(x)=-vF =. ‘Then, from the formula, the derivative of these functions is given by EXERCISES 4 Obtain J, and f, for lx, y) = x7 + 2xy + Sy 2 Under the gravitational acceleration g, the period T of a pendulum hav- ing length L is given by reaefE (the gravitational acceleration g is known to vary depending on the height from the ground). ‘Obtain the expression for total differential of T. IL is elongated by 1 percent and g decreases by 2 percent, about what percentage does T increase? 3. Using the chain rule, calculate the differential formula of the implicit function h(x) of fix, y) = cin a different way than above. 218 CHAPTER 6 LET'S LEARN ABOUT PARTIAL DIFFERENTIATION!EPILOGUE: WHAT IS MATHEMATICS FOR?WHO IS IN CHARGE OF THis OFFICE? eee WHAT 15 MATHEMATICS FOR? 221EXCUSE ME, DO. S (OH, HE IS ALWAYS YOU KNOW WHERE ‘SWIMMING. ‘THE PERSON IN \ CHARGE 1S? THERE YOU ARE!If i VY LA MR. SEKI I DECIDED TO SPEND ONE MORE YEAR THINKING ABOUT THINGS IN ‘A WARM PLACE.TO CONVEY THINGS THAT CANNOT BE TOMORROW WiLL BE ANOTHER GREAT DAY. WHAT ARE WEA SOLUTIONS TO EXERCISES PROLOGUE 1. Substituting 5 35 = 3 (x-92) inz=7y-30,2= 8 (x -32)-30 y= S(x-92)imz=7y (x -82) CHAPTER 1 + A. f(6) = g(5)= 50 S15) =8 2 tg FA*)~ $10) _ yy (08) 22 ‘Thus, the derivative of fix) is ftx) = 3x7. CHAPTER 2 1. The solution is S(x)=-2 f'(x)=8x*-12=8(x-2)(x+2) When x < ~ 2, Fe) > 0, when -2 2, f(x) 3. Since fix) = (1 ~ 2)’ is a function g(h(x)) combining g(x) = x° and hix) = Tex. F(x) =9'(h(x))h’(x)=3(1-x)*(-4)=-801-x) 4. Differentiating g(x) = x1 - x)° gives a(x)=(#) 0s) +8 (0-2) =2x(1-x)’ +x*(-3(1-x)') = x(1—x)' (2(1-x)-3x) =x(1-x)' (2-5x) g'(x)=0 when x2 or x=1, and g(1)=0. 108 3, 2 aiz5 * *"5 ‘Thus it has the maximum g| 2 CHAPTER 3 1. The solutions are © f'sx’ax=x'|'=3° 1 =26 © [fxs (iex)'des [2x-(14x*) de = [Pande =5* -0° =25 226 APFENDKA2 A. The area between the graph of y = fix) x? Sxand the xaxis equals Since x) et +x(e") =e% + xe" =(1+x)e" the minimum is Setting fix) = x? and g(x) = In x, integrate by parts. Jf (=*) in xd + [58 (In xy de =e? nein SOLUTIONS TO exexcises 227Thus, [f2xinsae sfx? Lae e" ffoetnade =f" xd +e? Bie CHAPTER 5 1. For S(x)=-e*, f(x) =e, 57"(; (0)=-1, $°(0)=1, $7(0)=-1... 2. Differentiate J(x)=(cosx)". f'(x)=(cosx)* sinx S(x)=2(cosx)* (sinx}’ +(cosx)* cosx =2(cosx)* (sin x}* +(cosx)* from f(0)=1,$'(0)=0,5"(0)=1 3 Proceed in exactly the same way as on page 155 by differentiating fix) repeatedly. Since you are centering the expansion around x = a, plugging in a will let you work out the ¢,s. You should get ¢, = Unt f(a), as shown in the formula on page 159. 228 APPEND ACHAPTER 6 1 For fx, y) =x" + 2xy + 3y°, f,= 2x + 2y, and f, = 2x + 6y. 2. The total differential of raan|E aang tt? Vo is given by ar, ar=¢ ag dL =-ng*Ldg+ ng *L db Thus. AT =-ng *L?Ag+ng *L?AL Substituting Ag «0.029, = 0012, we ge AT =0.02rg *L?g + 0.01xg *L?L = 00029"! ~003! «01st So T increases by 1.5%. 3. Ifwe suppose y = h(x) is the implicit function of flx, y) = ‘Thus, since the left side is a constant in this region, flx, h(x)) = ¢ near x. From the chain rule formula SOLUTIONS TO BxERcisES 229