0% found this document useful (0 votes)
61 views26 pagesCurve Sketching and Inequalities
Uploaded by
HENRY NEWTONCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
61 views26 pagesCurve Sketching and Inequalities
Uploaded by
HENRY NEWTONCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 26
CURVE SKETCHING and INEQUALITIES
Brailey Sims o
Department of Mathematics
‘The University of Newcastle
NSW 2308
The graph of a function js a powerful visual tool in analyzing and understanding the
function. ‘The adage ‘a picture is worth a thousand words’ is particularly relevant to the
mathematical sciences. Sketching graphs allows you to see the features and behavior of
functions and formulas.
‘The importance of graphs is not confined to mathematics. Graphs are used throughout the
physical sciences; physics, chemistry, biology and geology, the social sciences, engineering,
economics and business.
Sketching a function's graph should become second nature to you. It is a skill, like swim-
ming or driving, which once mastered will come naturally for the rest of your life. Sketching
a function's graph should not be seen as the culmination of a tedious sequence of evalu:
ations, differentiations, and limits, but rather as the first step in any skirmish with the
function.
In these notes I want to introduce and explore a systematic approach to the graphing
of functions which is both powerful and efficient. If you were asked to differentiate the
function. :
you would think of it as the quotient
i)
gz)"
of the two ‘simpler’ functions f(z) = 2 and g(z) = 1, and proceed to build its
derivative using those for f(x) and 9(2). A similar approach can be used when it comes to
graphing such a function. We would first sketch graphs for f(z) and (x) and then begin
to build a graph for the quotient from them. As a preliminary, illustrate the following points on the numberline given below:
a+, a—b, ab, a/b, 1/a, Ya, Vb, a”, B, b
4
NOTE: The numberline has been drawn vertically because we will usually want to operate
2 with ordinates of graphs.
POWER FUNCTIONS
‘An important class of functions are the power functions of the form y= Ac®.
‘Typical graphs of power functions for « > 0 and both A and a positive are sketched below
y= 2° for a= 1/3, 1/2, 1, 2 and 3
‘To understand these graphs observe:
(1) For all values of A and a the graph of y = Ax® passes through the point (1, A).
(2) dy/dz = aAz*~, so for A and a positive the functions is increasing. For a greater
than one the slope increases as x does, while for a between zero and one it decreases.
(3) In order to compare the graphs for different note that for z greater than one the
\pé are
graphs for larger « lie above those with smaller a. For z between zero and one, the
reverse is the case; that is, graphs for larger @ lie below those with smaller a.
When A is positive and @ is negative, the function y = Az® is decreasing. When z = 0
the function is undefined with the y-axis as a vertical asymptote, Some typical examples
are sketched below: 1
y= for @=~1/2, —1 and ~2
‘The graphs with A negative are the reflections in the z-axis of those for A positive:
a
=2* for a =—1, 1/2 and 2
Exercises;
(1) Sketch the graph of y= 2° fora=,a=2,a=
(2) Ha is increased from 2 to 10, how does the graph of y = 2° change?
4 POLYNOMIALS
Linear functions, quadratic functions and power functions with whole number exponents
are all special cases of the general polynomial function
y= p(t) = ag + at + age? +--+ + anz”
‘We consider some general features which help in sketching the graph of a polynomial.
Asymptotic Behaviour: When the magnitude of z is large, the value of p(x) is dominated
by the highest degree term a2”. Thus the larger the magnitude of x, the more nearly the
graph of y = p(2) looks like that of the power function y = anz™.
‘The following diagram shows the different kinds of behaviour that can arise.
a, positive and
n even
or a, positive
a, negative and
n odd
a, positive and
n odd 7
a, negative
or n
a, negative and
nm even
For example; when p(z)
— 32? + 22, we have the following behaviour:
Roots:: A root of the polynomial y = p(z) is a value 2» of z for whigh p(=) = 0. Thus ro
is a root of the polynomial precisely when the graph of y = p(z) intersects the z-axis at
5 r=2y.
A polynomial of degree n can have up to m roots. In the case of quadratics the roots are
given by the quadratic formula, More complicated formulas exist for the roots of cubice
and quarties (polynomials of degree 4). For polynomials of degree 5 and more it has bees
Proved that no general formula for the roots exists. Tn general finding roots canbe quite
difficult,
When the polynomial has integer coefficients it is sometimes possible to locate a root by
trying values of such as 0, 1, ~1, 2, or simple fractions such as 1/2, —2/3, ete. Here it is
important to remember that when expressed with lowest denominator, the numerator of
any rational root of p(x) must be a factor of the constant term %, while the denominator
of such a root must be a factor of the highest order coefficient an.
For example. When p(z) = 2? — 32? 4. 22:
for 2 = 0, we have p(0) = 0, s0 0 is a root,
for 2 = 1, we have p(1) =0, so 1 is a root,
for 2 = —1, we have p(—1) = —2, so -1 is not a root,
for = 2, we have p(2) = 0, so 2 is a root.
Since p(x) is a cubic and so can have at most 3 roots, we conclude that the roots are 0, 1
and 2. Combining this with the asymptotic behaviour noted above, we see that thie graph
of y= 2° — 32? + 22 looks like
As this exeimple illustrates, betwéen ainy two coiiseaiitive réots of a polynomial, we have
either at least one local maximum or at least one local minimum. Another useful obser
Vation is that ifthe values of p(2) at two points 2 and xy are of opposite sign, then p()
has a root between 21 and x2.
For example. If y = 82° — 132? ~ 112 +10, we have (0) = 10 and p(1) = —6 which are of opposite sign, so there is a root between 0 and 1; p(~1) = 0, so =1 is a root; p(2) =
so 2 is a root.
The graph will therefore look: like
positive
negative
Graph of y = 825 — 182 — 112 + 10
Provided we have enough information about the roots, it is usually possible to sketch the
graph of a polynomial.
For example. If y = p(2) = 42° — 8s? + 5x — 1, we find by substituting values for that
== 1 is root. Factorizing gives
y=(2—- (42? — 42 +1)
=(2-1)4(2? —2 +1/4)
= A(z —1)(z — 1/2)?
and so in this case p(x) hes roots at x = 1 and x = 1/2, with the root at ¢ = 1/2 being
repeated. Since these are the only points where the graph can cut the z-axis, we are able
to infer from the asymptotic behaviour that the graph must look like one of the following
possibilities: “hat it is indeed the heavy curve and not the broken one is easily confirmed by noting
that (3/4) = -1/16 is negative.
From the observation that a repeated root is also a root of the derivative we are able to
see that at a root which is repeated an even number of times, a polynomial will have a
local maximum or minimum and the graph will touch the z-axis without crossing it, as
illustrated in the last example. At a root which is repeated an odd number of times, the
graph crosses the z- axis, but has horizontal slope at the root.
For example. y = 24 — 23 = 23(2 — 1) has roots at 0 and 1, with the root at 0 being
repeated 3 times. The graph looks like
When the roots are ditticult to find, or provide us with insufficient 4
seck alternative methods, such as locating the local maxima and mi
For example. The polynomial y = 2° += is easily seen to have only one real root, at
«=0. Its graph cannot be sketched with certainty using the above methods, but noting
that dy/de = 32? +1 is strictly positive allows us to conclude that the polynomial is a
strictly increasing function and so its graph looks like Exercise For each of the following polynomials p(x), sketch the graph of y = p(z):
(a) plz) = (@ - 1) — 2)(z - 3)
(b) plz) = (@ +1)(2 - 2)?
(c) p(t) =2? —22-3
(@ a
(e) ef —223 422-1
=3¢-
TRIGONOMETRIC FUNCTIONS
In order to sketch the functions sine (and cosine) it is enous that you understand the
following construction.
Oher trigonometric functions can be sketched using identities like
cos(z) = sine + x/2)
see(z at
sin(2)
tan(z) = costa)
asin(c) + beos(2) = Asin(x + 4)
where A= Va? +B, and ¢ is such that cos($) = a/A, and sin(g) = 0/A
and the techniques we are about to describe.
9 SIMPLE COMPOSITES
Suppose we know the graph of a function y = f(z), for example;
y
The graph of y = f(z) = xsin(2)
then:
For ca constant the graph of y = j(2)+ciis that of f(2), but shifted vertically upward
a distance ¢, Thus for f(x) =zsin(z) and c= 2 we have
Graph of y= f(z) +2=asin(z) + 2
Of course, when < 0 shifting vertically upward a distance c amounts to a downward
shift of |o].
For ¢ a constant the graph of y = cf(x) is that of f(2), but with vertical distances
sealed by ¢. When ¢ < 0 this involves a scaling of verticle distances by || and a
reflection in the c-axis. Thus for f(x) given above and c= —1/2 we have
10 Graph of y= —2 f(z) = ~§ sin(x)
For ¢ a constant the graph of y = f(z +c) is that of f(z), but shifted horizontally a
distance c to the left. When c < 0 this means a horizontal shift of |c| to the right.
‘Thus for f() given above and ¢ = 2 we have
y
Graph of y = f(x +2) = (2 +2)sin(x + 2)
For ca constant in the graph of y = f(cz) as we move along the z-axis ‘things’ happen
c times as quickly as they did for y = f(z). Thus for f(z) given above and c = 2 we
have
Graph of y = f(2z)
2z sin(2z)
cy When ¢ <0 we obtain the graph of y = f(|clz) reflected about the y-axis,
‘The graph of y = ./f(z) is only defined at those x for which f(z) > 0 and at such
points may be obtained from the graph of y = f(z) by plotting an approximate square
toot of the ordinate, Here it is useful to remember the shape of the square root graph
(see the section on power functions), in particular recall that for 0 1 we have 1 < fa 0, and is the
reflection in the z-axis of y = f(z) where f(z) < 0. ‘Thus the graph lies entirely above
the a-axis. For f(x) given above we have
12 a x
Graph of y = |f(z)| = |zsin(=)|
Since y? = f(z) implies that y = t/f(z) the graph of y? = f(x) is that of y = f(z)
togethier with its reflection in the a-axis. Thus for f(z) given above we have
Y.
ste
Graph of y? = f(z) = zsin(z)
Remark; Functions such as 2/f(s) +1 which are built up from a sequence of simple
operations on f(z) may be graphed by successively constructing the appropriate sequence
Exercise Sketch y = 2/F(z) +1 when f(x
of graphs, in this case the graph of y = f(z) then y = f(z) +1 from which we can sketch
Vi (2) +1 and lastly y = 2/f(e) +1.
sit
RATIONAL COMBINATIONS
We will see that the task of graphing a complicated function is often made easier if we
recognise the function as a combination of simpler functions whos graphs are known or
readily found. ‘This extends the process already started in the last section.
Sums of Functions: Given two functions f and g, a graph of their sum y = f(x) + 9(z)
may be construcied from graphs of f and g (preferably drawn on the same axes) by
noting that the vertical distance (ordinate) fom a point z on the z-axis to the graph of
13 y = f(z) +9(2) is the sum of the ordinate at z to the graph y = (x) and the ordinate at 2
to the graph of y = (2). Thus, in the illustration below the distance AD = AB+AC. The
addition of these distances can be done in many ways, for example; by “eye”, by marking
them on the edge of a piece of paper, by using a pair of dividers (or your fingers as dividers),
or by using a ruler. You should aim to become practised in doing this. Particular care
should be taken at points where the graph of one or other of the functions y =-f(2) and
¥ = (2) lies below the z-axis and so the corresponding vertical distance is negative.
For example. Graps of the functions y = f(x) = 52(z ~ 2) and y = g(z) = 152 — 30
together with the graph of y = f(2) + 9(z) constructed from them are sketched below.
y = £(x) + g¢x)
= s(x)
a
The difference f(z) — 9(2) can be constructed similarly, by taking away rather than adding
the appropriate vertical distances. Altematively, it may be obtained by frst sketch y
~1g(z) and then adding this to the graph of y = f(z). Again, special care should be talen
when one or other of the graphs is below the z-axis.
Products: From graphs of the two functions f and g we can sketch a graph of their product
u = f(2)9(2) by performing ‘approximate’ multiplications (mentally, or otherwise) of f(z)
and g(z). In doing this use should be made of simple observations such es: the product of
two numbers larger than 1 is bigger than either of them; the product of a number larger
than 1 and a positive number less than 1 lies between them; the product of two positive
numbers smaller than 1 is less than either of them; the product of a positive number and
a negative number is negative.
For example. From graphs of y= 1-2 and y = /Z, we see that y
(1-2) VF has the
graph shown below:
14 Quotients: From graphs of the two functions f and g we can prepare a graph of the
quotient y = g(z)/f(z) by approximately dividing the value of g at 2 by the value of f at
2.
Alternatively the quotient 9(z)/f(z) can be regarded as the product of g(x) and the
reciprocal 1/f(z). ‘Thus it is often convenient to draw, as an intermediate step, the graph
of y = 1/f(z). To do this, it is necessary (mentally, or otherwise) to divide 1 by f(z)
Remember: taking the reciprocal of a smaller number gives a larger number, and vice
versa. In particular, if the values of f(z) are approaching 0 as x approaches xo, then, as t
approaches zo, the values of f(z) will tend to +00 or —co depending on whether f(x) is
positive or negative.
For example. To graph
we may first graph y =
and finally multiply this by z to obtain the desired graph. This is illustrated below. NOTE; for large 2, 1~ 2? behaves like ~2? and so 2/(1~ 22) is like 2/(—2”) =
‘Thus, as z becomes large in magnitude, z/(1 — 2?) approaches zero
Exercises
(2) Por the functions f(z) = 2 ~ 42” 452 ~2 and g(2) = 32 —2?, sketch graphs of f,
9, F +9, and fg on a common set of axes.
(2) In each of (a), (b) and (c) graph the functions indicated on a common set of axes,
(2)
(b)
@) y=i+2?
(i) y= 1/1 +2%)
(ii) y=2/(1 +22)
@
(i)
3)
(x)
16 (c) (@) y= f(z), where f(x) = 23 — 32? 422
(@ y=s@P
(GH) y= Vi)
INVERSE FUNCTIONS
Given a function y = f(z) we can ask; for what value of z does f(z) have a given value y?
In order that this question have an unambiguous answer it is necessary that each value y
of the function f come from only one value of z.
unambiguous
ambi guou:
answer, Le
answer
For a function to be unambiguous in this way each horizontal line must cut its graph only
once (just as each vertical line must only cut it once in order that f is a function). In any
region where the function is continuous this means it must be either strictly increasing or
strictly decreasing.
Such an unambiguous f is invertible and we can define the inverse function
z= f(y)
which assigns to each value y of f the unique x for which y = f(z). By reflecting this diagram about the line y = x so that the y-axis becomes horizontal and
the z-axis vertical we obtain a graph of the inverse function z = f(y).
NOTE; when y = f(z) is an invertible function these procedures give the graph for z =
£-"(y). If the variables do not have meaningful names which we would wish to preserve,
it is conventional to use z and not y for the independent variable. ‘To achieve this it i
necessary to swap the roles of ¢ and y in the final answer. Of course the functions given
by
i)
and
)
are of course the same function; only the names of the variables have been changed to
protect the convention.
Exercise For each of the following functions determine whether or not the function is
invertible. When the function is invertible draw graphs of both the function and its inverse
on the same set of axes. Find an expression for the inverse function using z to represent
the independent variable.
@ v= fa)=52~-3.
Gi) y= f(z) =
Gil) y= F(@) = (+1), 2 ¥-1.
Gy) v= ¢
2 +1, where 2 is positive.
18 INEQUALITIES
‘The general inequality F(z) < (2, <, >) G(z) can be converted to the form f(z) < (<) 0
by taking f(z) = F(x) — G(2), or f(z) = G(x) — F(z). Now, the graph of f(x) can only
cross the z-axis by either cutting it, or jumping over it; that is, at a zero of f(z), or at a
point of discontinuity for f(z). Between consecutive pairs of such points f(x) must remain
on one side of the a-axis, and so the inequality is either true, or false, throughout the whole
of such a segment. ‘To decide which it is therefore only necessary to test one ‘conveniently’
chosen point in the segment. Doing this for each segment and tesing each division point
between segments solves the inequality. It should be noted that if unnecessary division
points have been included the method, while requiring additional computation, still leads
to a solution.
Zeros of f(z) occur at precisely those points where the original inequality is an equality,
and for the type of functions with which we will be dealing discontinuities will always be
signalled by the vanishing of a denominator in one of the expressions. Thus we may turn
the above observations into an algorithm as follows.
‘To solve the inequality F(z) < (2, <, >) G(z):
‘L) Find those 2 at which equality occurs; that is, solve F(z) = G(z), and plot these
ql
points on the numberline.
(2) Determine those z at which a denominator in one of the expressions involved in
F(z) or G(z) vanishes. Plot all such points on the numberline.
ther
(3) The points identified in (1) and (2) divide the numberline into a number of intervals.
From each of these intervals choose a convenient value of s and substitute it into the
inequality to test whether or not the inequality is true throughout that interval. Also
check whether or not the inequality holds at each of the points from (1) and
obtain a complete solution.
For example. To solve the inequality
(2) Equality occurs when 1/(2 2)
z, that is when 1 = or equivalently when
0
a? -I24+1= (2-1)
‘Thus we have equality when # = 1.
(2) The denominator of the LHS vanishes when z = 2.
(3) From (1) and (2) the numberline is divided into the intervals:
—o 2, and false otherwise,
= > 0.
Exercise (3/4 Unit, 1990) Solve
What follows makes particular use of the observation that any unnecessary division points
introduced by the fret two steps are eliminated when the intervals and division points are
tested.
Expressions involving absolute values: When the expressions involved contains ab-
solute values; for example,
1
le—g24
=
it is often more economical to modify the above procedure in the way illustrated below.
For any value of z the expression [2 — 2| is equal to r—2 or —(z—2). Consequently the
values of » for which equality occurs; that is, for which
1
|z-2|= :
will be included among those » which satisfy either ¢— 2 = 2 or —(p —2) = 2; namely
#= 1 (the root of 2?—22-+1 = 0) and 2 = 1—/2,z =14,/9 (the roots of 2? 22-1 = 0).
we use these 2, together with any values of z at which a denominator might vanish (shat
is, a discontinuity might occur), in this case 2 = 0, to divide the line into intervals, then
after testing each of the points and intervals we must have the solution
20 In our example the line is divided as follows.
v yox x x yoo¥
ty
1-72 0 1
After testing, the results of which are represented by ./’s or x’s in the above diagram. we
see that the solution is 7
z1+Vv2.
Problems involving the absolute value of more than one expression, or absolute valus in the
expression for a denominator, may be treated similarly, except that we solve for equality,
or the vanishing of the denominator, with every possible choice of signs in front of the
expressions appearing inside absolute values.
For example. To find those values of x for which
{e+1|-[22-121
We find those x for which
(e+1)-(@2-1)=1, that is
(z+1)+@z-1)=1, that is
-(e +1)-(22-1)=1, that is
that is
(+1) +(22-1)=
‘These points divide the line as follows
Testing the points and regions we have the solution
1
S281
a Constrained regions in the plane: A procedure similar to that used above may be
used to find the region in the plane determined by a set of inequality constraints, Each
inequality is investigated for equality. When absolute values occur in an inequality we
replace it by several equalities, one for each choice of signs in front of the terms inside
absolute values. The resulting equations divide the plane into regions. Each region is then
tested against the corresponding inequality using a convenient point in it and shaded out
if the inequality fails to hold.
For example. To determine the region consisting of those points (z,y) which satisfy the
constraints
le-yl <1
y $%1—2")
z20
y20
we proceed as follows.
From the first constraint, |z ~ yl <1, we obtain the two equalities
g-y=l
and
or equivalently
and
ysrsh
‘These are equations for the two lines illustrated below, which divide the plane into three
regions labelled A, B and C.
‘To test each of the regions we may note that:
(0,2) is in A and jo — 2)
should be shaded out. ‘The origin (0,0) is in B and [0 ~ 0|=0 < 1, so points in B satisfy the inequality.
(0,2) is in C and [0 ~(~2)] = 2 £1, so C’ should be shaded out.
‘Thus the first inequality leads to the diagram and shading illustrated above.
Replacing the second constraint, y < 2(1— 2”) with an equality leads to the parabola
y = 2(1— 2”), illustrated below superimposed on our previous picture.
‘The parabola divides the plane into two regions.
‘Testing the region above the parabola using the point (0,3) we see that 3 ¢ 2(1—0°
and so this region must be shaded out, as has been done in the diagram above. On
other hand (0, 0) is in the region inside the parabola and 0 < 2(1 — 0?) = 2 so this region
receives no further shading.
Similarly, we may complete the picture by including the two constraints > 0,y > 0.
to obtain the situation depicted below in which the solution is the region left unshaded
together with its boundary along which the non-strict inequalities of the constraints are
satisfied.
23 Exercises
(1) For each of the following indicate on a number line th
jose points « which satisfy the
inequality, also describe the set of such z by means of
simple inequalities
(a) e@+1<32-2. 7
(b) 22-5) <4
(©) Pe-iset2.
(4) 22 ~1)-Je +4 21.
oO
gy S872 aes.
(2) For each of (a) and (b) indicate those points (
inequalities,
(a)
2, y) in the plane which satisfy the given
ttys4
ySde
2asy
3
