How much has built up in total when only the rate is known? The chapter on integrals starts with this question. That is the case with water in a tank and with the cost of an order. You can already find the rate from the volume: that is the derivative. The integral does the reverse and finds the volume from the rate.
An integral builds a total out of the rate at which it comes in. The rate can be many things:
cubic metres of water per hour;
euros of cost for each further item;
money per year into an account.
Water flows into an empty tank. At first the tap is almost closed, then it is opened wider and wider: after t hours the water flows at a rate of 2t m³ per hour.
How much water is in the tank after 4 hours?
Multiplying the rate by the time will not work: the rate is different at every moment.
Here, though, the speed grows evenly: from 0 at the start to 2⋅4=8 m³ per hour after 4 hours. The water in 4 hours is the area under the speed graph, a triangle with base 4 and height 8:
24⋅8==16 m3
After 4 hours the tank holds 16 m³. When the speed graph is curved, there is no triangle any more, and the integral finds the area under it.
Strips: rate times time
Start with the simple case: the water flows evenly, 3 m³ per hour. How much flows in during 4 hours?
A single multiplication is enough here: in 4 hours, 3⋅4=12 m³ flow in. If you draw the rate as a graph above the time axis, you get a rectangle of height 3 and width 4. The amount of water is its area.
In our tank the rate grows: f(t)=2t m³ per hour. You cannot multiply one rate by 4 hours, because the rate keeps changing. So you cut the time into short pieces. Within a short piece the rate hardly changes, and the water in that piece is roughly “rate × length of the piece”. In the picture this is the area of a narrow strip. All the water in 4 hours is roughly the sum of the strips.
Step 1 of 3Four strips of one hour
Cut the 4 hours into 4 pieces of 1 hour each. Take the height of each strip from the rate at the right edge of the piece: f(1)=2, f(2)=4, f(3)=6, f(4)=8. Each is 1 wide, so the sum is 2+4+6+8=20 m³. The strips stick out above the line: inside each piece the rate was lower than at its right edge. So 20 is more than the true amount. The right edge is only a convention: the left edge would give 0+2+4+6=12 (14 with eight strips), which is less than the true amount; with very thin strips both sums approach the same number.
Cut into 8 pieces of 0.5 hours each. The heights at the right edges are f(0.5)=1, f(1)=2, and so on up to f(4)=8. The heights add up to 1+2+…+8=36, each strip is 0.5 wide, together 0.5⋅36=18 m³. The strips are thinner, stick out less, and the sum is closer to the truth.
With a width of 0.25 hours the sum is already 17. The thinner the strips, the closer the sum comes to the area under the line itself. Under the line lies a triangle with base 4 and height f(4)=8; its area is 24⋅8=16 m³. That is the water in the tank.
The number that the sums of strips approach is called the integral (“Integral”) of f over the interval from 0 to 4:
∫042tdt=16
How to read the notation:
the sign ∫ is a stretched letter S, from “sum”;
the numbers at the bottom and top, 0 and 4, are the bounds of integration (“Grenzen”): the start and end of the interval;
the function under the sign, here 2t, is called the integrand (“Integrand”);
dt is the width of a very thin strip in t (the strip itself is 2tdt: height 2t times width dt); the letter after d says which variable you cut along.
The letter of the variable does not matter. ∫042tdt and ∫042xdx are the same number 16. Both letters, t and x, appear below.
If the tank already held 5 m³ at the start, after 4 hours it holds 5+16=21 m³. Every task of the kind “given a rate, find an amount” works this way: the starting value plus the integral of the rate.
Antiderivative: the derivative backwards
Cutting into strips every time takes long. The short way goes through the derivative. Here is a reminder of what it is.
The derivativeF′(x) shows how fast the function F changes near the point x: how much F grows over a small step in x, per unit of the step. If F is the volume of water, then F′ is the inflow rate. The article “What the derivative is: slope and the rules” derives this in detail. Here you need five of its rules:
power:(xα)′=αxα−1 for any number α: the exponent comes to the front and drops by 1, for example (x3)′=3x2, (t2)′=2t;
number: a number has derivative 0, and 5x has derivative 5;
factor and sum: a factor stays, (6x2)′=6⋅2x=12x; a sum is differentiated term by term;
exponential:e≈2.718 is a special number, the ex key on the calculator. The function ex is its own derivative: (ex)′=ex;
logarithm:lnx answers the question of which power you must raise e to in order to get x; for example, ln1=0, because e0=1. Its derivative is (lnx)′=x1. On the calculator this is the ln key, not log.
Back to the tank. Let V(t) be the volume of water that has flowed in by time t. The rate at which V changes is the inflow rate, so V′(t)=2t. So we are looking for a function whose derivative is 2t. The power rule gives the answer: (t2)′=2t, so V(t)=t2. Such a function is called an antiderivative (“Stammfunktion”):
F is an antiderivative of f if F′(x)==f(x)
For f(t)=2t, F(t)=t2 works. So do t2+5 and t2−100: the number at the end disappears when you differentiate. That is why one writes F(t)=t2+C, where C is any number.
You can always check an antiderivative you have found: take its derivative (this is called “differentiating”), and you must get f.
Check yourself. Is F(x)=x3+7 an antiderivative of f(x)=3x2? And F(x)=6x?
The fundamental theorem: F(b) − F(a)
Back to the tank with rate 2t, and let us ask differently: how much water flowed in between times 1 and 4?
You can find it in two ways, and both give the same number.
Step 1 of 3Strips from 1 to 4
The water in this time is the area under the rate 2t from 1 to 4, that is, the integral ∫142tdt. Under the line lies a trapezoid: heights f(1)=2 and f(4)=8, width 3, area 22+8⋅3=15 m³.
The antiderivative V(t)=t2 tells how much water is in the tank by time t. By time 4 the tank holds V(4)=42=16 m³; by time 1 it held V(1)=12=1 m³.
What flowed in is V(4)−V(1)=16−1=15 m³. That is the same area 15: to find the area you do not need to cut strips, two values of the antiderivative are enough.
The same holds for any rate f and its antiderivative F: the amount from a to b equals how much F grew over that time. The reason: over a short piece F grows by “rate × width”, which is the area of the strip, and the growths of all the pieces add up to the whole growth.
∫abf(x)dx==[F(x)]ab==F(b)−−F(a)
This is the fundamental theorem of calculus (“Hauptsatz”). The notation [F(x)]ab means the same as F(b)−F(a). The number a is called the lower bound, b the upper bound.
It does not matter which antiderivative you take. The number C is in both F(b) and F(a) and cancels in the difference: (16+C)−(1+C)=15. The simplest is to take C=0.
Take ∫133x2dx.
Antiderivative F(x)=x3, because (x3)′=3x2.
Upper bound: F(3)=27.
Lower bound: F(1)=1.
Difference: 27−1=26.
Check yourself. What is ∫032xdx? And ∫232xdx?
Table of antiderivatives
For the rate 2t the power rule gave away the antiderivative t2. So that you do not have to guess afresh each time, the antiderivatives of common functions are collected in a table. Each row of the table comes from reading a derivative rule backwards. Example: which function gives x2 when differentiated? The derivative has an exponent 1 lower, so look among x3. But (x3)′=3x2, three times too much. Divide by 3: (3x3)′=x2. So the antiderivative of x2 is 3x3: the exponent goes up by 1, and you divide by the new exponent. The same goes for all powers:
Function f(x)
Antiderivative F(x)
Check: F′(x)
a number, for example 4
4x
4
xα, α=−1
α+1xα+1
α+1(α+1)xα=xα
x1, for x>0
lnx
x1
ex
ex
ex
The row for x1 is a separate one. The notation x−n means the same as xn1: this is how you write a fraction with a power of x below. So x1=x−1, and the power rule would make you divide by −1+1=0, but you cannot divide by 0. Instead, the derivative rules tell you that (lnx)′=x1.
The other fractions xnc (except n=1) are ordinary powers: xnc=cx−n, and the row for xα works for negative α too. For example, x24=4x−2. Two moves: the new exponent is −2+1=−1; then divide by it: 4⋅−1x−1=−4x−1=−x4. A root is a power too: x=x1/2.
Two more rules, as for the derivative. Why they hold can be seen with the tank:
Sum. Suppose two pipes fill the tank at once: the water from both pipes is the water from the first plus the water from the second. That is why a sum is integrated term by term, each term by its own row.
Factor. A pipe that flows three times as strongly gives three times as much water in the same time. That is why a factor stays in place.
For f(x)=6x2+4: 6x2 gives 6⋅3x3=2x3, the number 4 gives 4x, together F(x)=2x3+4x. Check: (2x3+4x)′=6x2+4.
Now you can see why F(0) is not always 0. For x3 it is, for ex it is not: e0=1. That is why in ∫01exdx=e1−e0=2.7183−1=1.7183 the one is subtracted.
Definite integral: compute it from a to b
Exam-style task
Integrieren Sie f(x)=x21+3 über dem Intervall [1;4].
a)1.97
b)8.25
c)10.63
d)9.75
e)10.72
The task on the sheet above asks you to integrate f(x)=x21+3 over the interval from 1 to 4, that is, to find ∫14(x21+3)dx. Signal: “Integrate” and an interval with two numbers, and under the integral a sum of powers, numbers and fractions. You need a definite integral: it has bounds, so the result is one number, not a function. “Integral of a sum” means that under the sign there is a sum of terms.
Why can you take the terms one at a time? Think of the tank with two pipes flowing into it at once: the water from both pipes is the water from the first plus the water from the second. The same with the integral of a sum: each term is integrated by its row of the table, the antiderivatives are added, and then you compute F(b)−F(a) once.
Step 1 of 4Write the fraction as a power
The table has no row for x21, but it has one for a power. x21=x−2, the exponent is α=−2.
x−2: exponent −2+1=−1, divide by −1: −1x−1=−x1. The number 3 gives 3x. Together F(x)=−x1+3x. Check: (−x−1)′=x−2 and (3x)′=3.
F(4)−F(1)=11.75−2=9.75. Rough check: the curve's height is between 3 and 4, the width is 3, so the area is between 3⋅3=9 and 4⋅3=12. 9.75 fits.
Ask yourself: where does the minus in −x1 come from if everything under the integral is positive?
This time there is a fraction under the integral. Find ∫13xx2+4xdx. The table has no row for a fraction with a sum on top, so you first turn it into a sum of powers.
Each term of the numerator is divided by x separately: xx2+x4x=x+4.
x gives 2x2, the number 4 gives 4x: F(x)=2x2+4x.
You do the same with a squared bracket: (x+1)2=x2+2x+1, then term by term. And with a fraction that cancels: x−5x2−25=x−5(x−5)(x+5)=x+5.
The sum under the integral contains a fraction x24, and its sign easily slips. Compute ∫13(3x2+2+x24)dx, rounded to two decimal places.
Term by term: 3x2→x3, 2→2x, x24=4x−2→4⋅−1x−1=−x4. Together F(x)=x3+2x−x4.
Bounds: F(3)=27+6−34=31.6667, F(1)=1+2−4=−1.
Take the difference and round.
The antiderivative of x1 is lnx, and the calculator gives the values (the ln key, not log): ∫18x1dx=ln8−ln1=2.0794−0=2.0794.
Your turn2 of 2
Find the value of the integral I=∫28(x6+3)dx. Give the result rounded to two decimal places.
Traps of the integral of a sum
Taking the derivative instead of the antiderivative. For ∫14(x21+3)dx this gives 1.97: the derivative of x21+3, −x32, was taken and the bounds put in. “Integrate” and “differentiate” are easy to mix up, and both work on the exponent. It is easy to catch: differentiate your F, and the integrand must come back.
Losing the minus when dividing by a negative exponent. For x−2 the new exponent is −1, and the antiderivative is −x1. Without the minus you get x1+3x and the answer 8.25. Sign: for a fraction of the form x2c the antiderivative has the opposite sign to the numerator: x24 gives −x4, and −x24 gives +x4.
Leaving a number as a number. A number under the integral gives x with that number as a factor. In the integral ∫14(3x2−4x−5)dx from “Check yourself”, −5 stayed −5 and cancelled in the difference: 33 instead of 18. The check catches it: the derivative of a number is 0, not −5.
Forgetting the lower bound. In the exercise above, F(8) alone gives 36.48. You always subtract F(a).
Integrating the numerator and denominator of a fraction separately. For xx2+4x this gives ∫13xdx∫13(x2+4x)dx=424.67=6.17 instead of 12. The integral of a fraction is not the fraction of the two integrals: simplify the fraction to a sum of powers first and integrate that.
Antiderivative plus a constant: K′(x) and the cost at one output
The tank that held 5 m³ from the start has already shown that sometimes a known value is added to the integral. The cost task is built the same way, except that the cost is known not at zero but at 5 pieces.
What is known: 5 items cost 2150 €, and the marginal costs are K′(x)=6+8x+9x2. Two things are asked:
What do 9 items cost?
What are the fixed costs, that is, the costs that do not depend on output?
Signal: you are given the marginal costs K′(x) and the cost at one output, and asked for the cost at another.
A workshop knows its marginal costsK′(x): how much more output costs with each further item when output is already x items (“Marginal cost and elasticity: what they mean”). K(x) is the cost of x items, and K′(x) is the rate at which it grows. It is the same tank: money instead of water, items instead of hours. The cost at the new output equals the known cost plus the integral of the rate:
K(b)=K(a)+∫abK′(x)dx
The marginal costs alone are not enough: they say how much more each further item costs, but not the costs themselves. The 2150 € alone are not enough either: they belong to 5 items, not 9. The answer comes only from using both together.
Step 1 of 5Antiderivative of the marginal costs
Term by term: 6→6x, 8x→8⋅2x2=4x2, 9x2→9⋅3x3=3x3. Together F(x)=6x+4x2+3x3.
F(9)=54+324+2187=2565. F(5)=30+100+375=505.
∫59K′(x)dx=2565−505=2060 €. That is how much more 9 items cost than 5.
K(9)=K(5)+2060=2150+2060=4210 €.
The fixed costs (“Fixkosten”) are the costs at zero items, K(0). K is an antiderivative of K′, and two antiderivatives of the same function differ only by a number: all costs have the form K(x)=6x+4x2+3x3+C. From K(5)=505+C=2150 we get C=1645, and K(0)=0+C=1645 €. Here the number C is no longer “any”: the known point K(5) fixes it.
Now a different order: 3 items cost 500 €, and the marginal costs are K′(x)=10+4x+6x2.
What do 7 items cost?
The cost is known only for 3 items. To get to 7 items, the increase in between is still missing.
Term by term: 10→10x, 4x→4⋅2x2=2x2, 6x2→6⋅3x3=2x3. Together F(x)=10x+2x2+2x3.
F(7)=70+98+686=854. F(3)=30+18+54=102.
∫37K′(x)dx=854−102=752 €. That is how much more 7 items cost than 3.
K(7)=K(3)+752=500+752=1252 €. Rough check: costs grow with output, and 1252 is more than 500.
The marginal costs are K′(x)=0.06x2−2x+30, and at 10 items the costs are 1000 €.
What are the fixed costs?
The 1000 € themselves are not the answer: they include a part that grows with output.
Antiderivative: K(x)=0.02x3−x2+30x+C.
At 10 items: 0.02⋅1000−100+300+C=220+C.
Set this equal to 1000 and find C=K(0).
Cost traps
The numbers here come from the example with 5 and 9 items: K(5)=2150, F(9)=2565, F(5)=505, correct answer K(9)=4210.
Answering with the increase alone.∫59K′(x)dx=2060 is easy to take for the answer: it is the last number computed. But 2060 is only how much more 9 items cost than 5. Check: the cost of 9 items cannot be less than that of 5, yet 2060<2150.
Forgetting the constant. Taking F(9)=2565 as the cost of 9 items. The temptation: the antiderivative is the cost function, so just put in 9. But F was chosen with C=0, and the real costs contain the fixed part 1645. Check: put 5 into your formula; it must give the given 2150, but F(5)=505.
Adding F(9) to the known costs without subtracting F(5).2150+2565=4715. The temptation: “known plus antiderivative” looks like the formula K(b)=K(a)+∫abK′(x)dx. But the integral is the difference F(9)−F(5), not F(9) alone. The extra 505 is exactly F(5).
Answering with the fixed costs.C=1645 is the cost at zero items. The question asks for the cost at 9 items. The temptation: the constant was found last and looks like the result. Check: reread at which output the cost is asked for.
Taking the given costs for the fixed costs. Answering 1000 € in the example where 10 items cost 1000 €. That is the total cost at 10 items, which already includes 220 € that depend on output. The fixed costs are K(0), and when costs grow, K(0) is less than K(10).
On the sheet:∫xαdx=α+1xα+1 for α=−1, ∫exdx=ex, ∫x1dx=lnx. The fundamental theorem F(b)−F(a) and the cost formula K(b)=K(a)+∫abK′(x)dx you have to remember. The formula sheet (“Formelsammlung”) is printed in the task booklet after the tasks.
Formula sheet and what to memorise
The sheet repeats the table of antiderivatives almost in full. The fundamental theorem and the cost formula are not on it.
On the formula sheet:
the basic integrals ∫xαdx=α+1xα+1(α=−1), ∫exdx=ex, ∫x1dx=lnx: they give the antiderivative for the integral of a sum and for the costs.
Memorise yourself:
the fundamental theorem ∫abf(x)dx=F(b)−F(a); F(a) is subtracted even when a=0;
costs from marginal costs: K(b)=K(a)+∫abK′(x)dx; fixed costs K(0)=C;
an antiderivative is checked by differentiating; the number at the end does not matter.
In short
The topic in a minute
The integral∫abf(x)dx is the sum of thin strips “height f(x) × width dx” from a to b: from a rate it gives the total amount.
An antiderivativeF of f: a function with F′=f. Any number C can be added to it.
Fundamental theorem:∫abf(x)dx=F(b)−F(a).
Table:xα→α+1xα+1, number → number ⋅x, x1→lnx, ex→ex; a factor stays, a sum is taken term by term.
“Integrate” with an interval, a sum of powers, numbers and fractions under the integral; or costs from marginal costs. Each term by its row of the table, then F(b)−F(a).
∫abf(x)dx==F(b)−−F(a)
Trap:taking the derivative instead of the antiderivative; losing the minus for x24; forgetting F(a); for costs, answering with the increase without K(a).