Use rational exponents and the laws of logarithms, change the base of a logarithm, and solve exponential and logarithmic equations, rejecting any solution that makes a logarithm undefined.
IB Mathematics: Analysis and Approaches
Part of 1 Number and algebra
SL and HL
Key terms introduced here
Each is in the Maths AA glossary, with every other term in the subject.
126 exam-style questions on this subtopic, written in the style of the IB papers. The first has its markscheme open. Counted on
Question 1 · AA-00449HL only · 8 marks
In a factory, 3% of the turbine blades produced have a hidden crack. Each blade is checked with an ultrasound test. If a blade has a crack, the test gives a positive result with probability 0.92. If a blade has no crack, the test gives a positive result with probability 0.06.
When a blade is tested more than once, the results of the tests are independent of each other, given whether or not the blade has a crack.
(a)
A blade is tested once and the result is positive. Find the probability that the blade has a crack.
[3]3 marks
Markscheme (a)
P(positive)=0.03×0.92+0.97×0.06=0.0858
Try a quick-fire question
One of 42 short multiple-choice questions on 1.7. Pick an answer to see why it's right or wrong.
Which is equal to loga+logb?
From the formula booklet
The entries of the mathematics: analysis and approaches formula booklet that 1.7 uses. Each opens in the explained booklet.
An equation in which the unknown appears in an exponent, such as 2x+1=5. It is solved by writing both sides with the same base or by taking logarithms of both sides.
A value produced while solving an equation, for example after squaring both sides or splitting into cases, that does not satisfy the original equation and must be rejected.
A number which, raised to the power n, gives a; written na=a1/n. For example 38=2 because 23=8. For even n the symbol means the non-negative root of a≥0.
An exponent that is a fraction nm. It means a root followed by a power: am/n=(na)m=nam, so 82/3=22=4.
Use the change of base formula to evaluate logarithms and to solve equations with logarithms in different bases.
Solve exponential equations by writing both sides with the same base or by taking logarithms.
Solve equations that become quadratics after a substitution, such as 4x−3×2x+2=0.
Solve logarithmic equations and reject any value that makes a logarithm undefined.
The laws of exponents in 1.5 were stated for integer exponents. Extending them to fractional exponents connects powers with roots, so that x can be written x1/2 and handled with the same rules as any other power. Every law of exponents also has a matching law of logarithms, which turns products into sums and powers into multiples. Together, these laws make it possible to solve any equation in which the unknown is an exponent, such as finding how long an investment takes to double or when a decaying quantity falls below a given level.
Rational exponents
Roots as powers
The laws of exponents from 1.5 can be used to decide what a fractional exponent should mean. If the power-of-a-power law (am)n=amn is to keep working, then
(a1/2)2=a21×2=a1=a
So a1/2 is a number whose square is a: it is the square root of a. In the same way, (a1/3)3=a, so a1/3 is the cube root of a. In general, a1/n is the nth root of a.
For example, 161/2=16=4, 81/3=38=2 and 811/4=481=3, since 34=81. When n is odd, negative numbers also have an nth root: (−8)1/3=−2, because (−2)3=−8. When n is even, a negative number has no real nth root, because no real number raised to an even power is negative.
Exponents of the form m/n
A rational exponent nm combines a root and a power. By the power-of-a-power law, am/n can be split in two ways:
am/n=(a1/n)m=(na)moram/n=(am)1/n=nam
The denominator n is the root and the numerator m is the power. Both orders give the same answer, but taking the root first keeps the numbers small: 272/3=(327)2=32=9, while 3272=3729 needs a less familiar cube root.
All the laws of exponents from 1.5 continue to hold for rational exponents when the base is positive. A negative rational exponent still means the reciprocal, so a−m/n=am/n1.
Exponent
Meaning
Example
a1/n
nth root
1251/3=5
am/n
nth root, then mth power
163/4=23=8
a−1/n
reciprocal of the nth root
36−1/2=61
a−m/n
reciprocal of am/n
8−2/3=41
a0.5, a0.25, …
write the decimal as a fraction
320.4=322/5=22=4
Simplifying expressions with roots
Expressions that mix roots and powers become easier to handle once every root is written as a power. Then the laws of exponents apply directly.
Writing roots as powers is also how expressions such as x1 are prepared for differentiation and integration in topic 5, where x1=x−1/2 fits the rules for powers of x.
The laws of logarithms
Logarithms undo powers
By the definition in 1.5, logax is the exponent to which a must be raised to give x, where a>0, a=1 and x>0. Two results follow at once. Raising a to that exponent gives back x, and the logarithm of a power of a is just its exponent:
alogax=xlogaax=x
The second gives the special cases loga1=logaa0=0 and logaa=logaa1=1. These four facts are used constantly when simplifying, so they are worth knowing by heart. For example, 10log7=7, eln5=5 and log33−4=−4.
The product, quotient and power laws
Since logarithms are exponents, each law of exponents gives a law of logarithms. Multiplying powers adds the exponents, so the logarithm of a product is the sum of the logarithms.
These three laws are in the formula booklet. Each one can be proved from the matching law of exponents.
Proof of the product law. Let p=logax and q=logay. In exponential form, x=ap and y=aq. Multiplying,
xy=ap×aq=ap+q
Rewriting this in logarithmic form gives logaxy=p+q=logax+logay.
Proof of the quotient law. With the same p and q, dividing gives yx=aqap=ap−q, so logayx=p−q=logax−logay.
Proof of the power law. With x=ap, raising to the power m gives xm=(ap)m=apm, so logaxm=pm=mlogax.
Law of exponents
Law of logarithms
ap×aq=ap+q
logaxy=logax+logay
ap÷aq=ap−q
logayx=logax−logay
(ap)m=apm
logaxm=mlogax
a0=1
loga1=0
a1=a
logaa=1
A useful special case of the power law is logax1=logax−1=−logax. The same laws hold for ln and for log (base 10), since these are logarithms with particular bases.
Combining and expanding logarithms
The laws work in both directions. Combining several logarithms into one is useful for evaluating expressions and for solving equations; expanding one logarithm into several is useful for simplifying and, later, for differentiation.
When combining, apply the power law first, so that every logarithm has coefficient 1, then use the product and quotient laws. When expanding, do the reverse: split products and quotients first, then bring down the powers.
A common question type gives the values of two logarithms as letters and asks for other logarithms in terms of them. The method is to write each new argument as a product, quotient or power of the given numbers.
Expressions with powers of e and 10
The rule alogax=x often appears in combination with the other laws. To simplify e2ln3, first use the power law to move the 2 inside: 2ln3=ln9, so e2ln3=eln9=9. Similarly, elnx−ln2=eln(x/2)=2x and 101+log5=101×10log5=50.
Change of base
The change of base formula
Calculators evaluate logarithms to base 10 and base e directly. To evaluate a logarithm in any other base, such as log210, it is rewritten in terms of one of these.
Proof. Let y=logax, so that ay=x. Take logarithms to base b of both sides and use the power law.
logbay=logbx⇒ylogba=logbx⇒y=logbalogbx
Since y=logax, this is the formula. Division by logba is allowed because a=1, so logba=0.
The change of base formula works with any new base b. Choosing b=10 or b=e gives a decimal value with a calculator; choosing a base that both numbers are powers of gives an exact value by hand.
Equations with logarithms in different bases
When an equation contains logarithms with different bases, change them all to one base before using the other laws. A base that the others are powers of is usually the best choice.
Solving exponential equations
An exponential equation has the unknown in an exponent. There are three main methods, and choosing the right one is the first step.
Form of the equation
Method
Both sides are powers of the same number, e.g. 9x=27x−1
Write both sides with the same base and equate the exponents
One power on each side with unrelated bases, e.g. 2x+1=5
Take logarithms of both sides and use the power law
A power and its square, e.g. 4x−3×2x+2=0
Substitute u=2x
Equal bases
If ap=aq with a>0 and a=1, then p=q, because each positive number has only one logarithm to base a. So when both sides can be written as powers of the same base, the exponents can be equated.
Taking logarithms of both sides
When the bases cannot be matched, take the logarithm of both sides. The power law then brings the unknown down from the exponent. Any base works; ln or log is usual because a calculator can evaluate them.
Solving for time in growth and decay
Exponential equations arise whenever a quantity is multiplied by the same factor in each time period, as in the geometric sequences of 1.3 and the financial models of 1.4. Logarithms find the number of periods. When the answer must be a whole number of periods, solve an inequality and round in the correct direction.
The doubling time in part (a) is close to 672=12. This is the rule of 72 used in finance: an amount growing at r% per year doubles in roughly r72 years. The rule works because t=ln(1+r/100)ln2 and, for small r, ln(1+r/100) is close to 100r, giving t≈r69.3; 72 is used instead because it has many factors and is slightly more accurate for typical rates.
y = 1.06ˣ
An amount growing by 6% per year doubles after about 11.9 years and triples after about 18.9 years, so it has more than tripled after 19 whole years.
When the growth factor is less than 1, as in decay or depreciation, its logarithm is negative. Dividing an inequality by a negative number reverses its direction.
Equations that become quadratics
An equation containing both a2x and ax is a quadratic in disguise, because a2x=(ax)2. Substituting u=ax turns it into an ordinary quadratic in u. After solving for u, each value is turned back into an equation for x. Since ax>0 for every x, any negative or zero value of u gives no solution.
The graph of y=4x−3×2x+2 confirms the two solutions of part (a). It crosses the x-axis exactly twice, at x=0 and x=1, and dips below the axis between them.
y = 4ˣ − 3 × 2ˣ + 2
The curve meets the x-axis only at x = 0 and x = 1. As x decreases, 4ˣ and 2ˣ approach 0, so the curve approaches the line y = 2.
Solving logarithmic equations
Rewriting in exponential form
A logarithmic equation has the unknown inside a logarithm. The usual method is to use the laws of logarithms to collect everything into a single logarithm, loga(…)=c, and then rewrite in exponential form: (…)=ac. If both sides are single logarithms with the same base, logaP=logaQ, then P=Q.
Checking the domain
Combining logarithms can produce values of x that do not satisfy the original equation. For example, logx+log(x−3) is defined only when x>0 and x>3, but the combined form logx(x−3) is also defined for negative x. Any value found by solving must therefore be substituted back into the original equation; a value that makes any argument zero or negative is rejected. Such a value is called an extraneous (or false) solution.
y = log x + log(x − 3)
The left-hand side of log x + log(x − 3) = 1 exists only for x > 3, so the graph meets y = 1 once, at x = 5. The value x = −2 from the quadratic lies outside this domain.
Logarithms on both sides and quadratic forms
When an equation has a multiple of a logarithm on one side, the power law moves the multiple inside first. Care is needed: 2lnx is defined only for x>0, even though lnx2 is defined for all x=0, so the domain must be taken from the original form.
Exam-style questions on exponents and logarithms
Questions on these laws appear on both papers. Without a calculator, answers are exact and often in a given form such as lnblna; with a GDC, the context is usually growth or decay, and the answer may be a whole number of time periods.
Further points that earn or lose marks:
The laws of logarithms and the change of base formula are in the formula booklet; the conditions a,x,y>0, a=1 are not always printed beside them.
An exact answer such as ln2ln5−1 is accepted in any equivalent form unless the question asks for a particular form.
When a whole number of periods is asked for, give the integer and show the inequality that leads to it.
Self-check
To do:
I can evaluate powers with rational exponents, such as 16−3/4, without a calculator.
To do:
I can write roots as powers and simplify expressions such as 3xxx.
To do:
I can prove the product, quotient and power laws of logarithms from the laws of exponents.
To do:
I can combine several logarithms into one and expand one logarithm into several.
To do:
I can use the change of base formula to evaluate a logarithm and to solve equations with different bases.
To do:
I can solve exponential equations by equating exponents, by taking logarithms, or by substituting u=ax.
To do:
I can reverse an inequality when dividing by the logarithm of a number between 0 and 1.
To do:
I can solve logarithmic equations and reject solutions that make an argument zero or negative.
A blade is tested n times and every result is positive. Find the least value of n for which the probability that the blade has a crack is greater than 0.999.
[3]3 marks
Markscheme (c)
0.03×0.92n+0.97×0.06n0.03×0.92n>0.999(M1)
EITHER
n=3 gives 0.991110… and n=4 gives 0.999415…(A1)
OR
n>3.80320…(A1)
THEN
n=4A1
Note: Accept an equation for the (M1).
Question 2 · AA-00460HL only · 28 marks
This question asks you to investigate how the probability that a component is faulty changes as the results of repeated tests are obtained.
A factory makes electronic components. Each component is either faulty (F) or good (F′). A test on a component gives one of two results, "fail" or "pass". The test has accuracy a, where 21<a<1: a faulty component fails the test with probability a, and a good component passes the test with probability a.
When a component is tested more than once, the results are independent of each other, given whether or not the component is faulty.
For any event E with 0<P(E)<1, the odds of E are defined as O(E)=P(E′)P(E)
(a)
Consider the case P(F)=0.1 and a=0.8. A component is tested once and fails. Find the probability that it is faulty.
Check: 2x+1=727>0 and x−1=73>0, so the solution is valid.
(c) This is a quadratic in log2x. Let u=log2x.
u2−3u+2=0⇒(u−1)(u−2)=0⇒u=1 or u=2
log2x=1 gives x=2, and log2x=2 gives x=4. Both are positive, so both are valid.
.
log39+2log3a−21log3b=2+2p−21q
(b) Combine with the product law, then rewrite in exponential form.
log2x(x+2)=3⇒x2+2x=8⇒(x+4)(x−2)=0
So x=2 or x=−4. Since log2x needs x>0, reject x=−4. The solution is x=2.
(c) Write 4x=(2x)2 and 2x+1=2×2x, then let u=2x.
u2−2u−8=0⇒(u−4)(u+2)=0⇒u=4 or u=−2
2x=4 gives x=2. 2x=−2 has no solution, since 2x>0. The solution is x=2.
Take natural logarithms: t=ln(1.08/0.95)ln4=10.808…≈10.8 years. The same value comes from the intersection of the two graphs on the GDC.
(b) A=500×1.0810.808…=1148.8…, so both populations are about 1150 fish.
(c) Solve 500×1.08t>1000, so 1.08t>2 and t>ln1.08ln2=9.006…. Population A first exceeds 1000 during the tenth year, so after 10 whole years (check: A(9)=999.5 and A(10)=1079.5).
(M1)
. Similarly,
O(E∣R)=P(E′∣R)P(E∣R)
for an event
R
.
.
[3]3 marks
(c)
Find a similar expression for O(F∣pass) in terms of a and O(F).
[2]2 marks
(d)(i)
Write down P(Rk∣F) and P(Rk∣F′).
[2]2 marks
(d)(ii)
Hence show that O(F∣Rk)=(1−aa)k×O(F).
[3]3 marks
(e)
A component is tested m+n times. It fails m of the tests and passes the other n, in some order. Let S be the event that these results are obtained, in that order. Show that O(F∣S)=(1−aa)m−n×O(F).
[3]3 marks
(f)
Explain briefly the significance of the result in part (e) for a component that fails exactly as many tests as it passes.
[2]2 marks
(g)
A component is tested 10 times. It fails 6 tests and passes 4. Find the probability that it is faulty.
[3]3 marks
(h)
A component is tested repeatedly and fails every test. Find the least number of tests after which the probability that it is faulty is greater than 0.999.
[3]3 marks
(i)
A faulty component is tested 10 times. Find the probability that, from its results, the probability that it is faulty is calculated to be greater than 0.5.
[4]4 marks
.
[2]2 marks
(b)
Write down the value of r, Pearson's product-moment correlation coefficient, for these data.
[1]1 mark
(c)
A scatter diagram suggests that the relationship between t and y is not linear. Let z=lny.
The regression line of z on t has equation z=ct+d. Find the value of c and the value of d.
[2]2 marks
(d)
Find the value of Pearson's product-moment correlation coefficient for t and z. Hence state which of the two regression lines gives the better model for the data.
[2]2 marks
(e)
Show that the equation z=ct+d can be written in the form y=Aekt, and find the value of A and the value of k.
[3]3 marks
(f)
Use the model y=Aekt to estimate the number of downloads in week 9.
[2]2 marks
(g)
Comment on the reliability of your estimate in part (f).
[1]1 mark
(h)
According to the model y=Aekt, find the first week in which the number of downloads is more than 100 thousand.