1a Maths Functions
1a Maths Functions
Young
1 Mathematical functions
2 Approximation by series
3 Indeterminate forms
Reference :
1
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
2
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
1 MATHEMATICAL FUNCTIONS
e x + e−x e x − e−x
cosh x = sinh x =
2 2
Not all functions can be represented analytically; indeed, any ‘reasonable’ curve you
care to sketch on a graph of f(x) versus x will define a function. For mathematical
calculations, such functions may have to be represented numerically at discrete points.
Alternatively, it may be possible to approximate a function over a certain range by using the
elementary functions, perhaps in some combination. The range over which the approximation
remains accurate may be quite large or quite small depending on how the original function
behaves. This idea of approximating a complicated function with a simpler mathematical
expression forms the basis for the following notes.
We know that x2 = (−x)2 and x4 = (−x)4 and so on for all even powers of x. We also
know that x = −(−x) and x3 = −(−x)3 and so on for all odd powers of x. Thus, graphs
involving even powers of x are symmetrical about the y-axis, and graphs involving odd
powers of x are anti-symmetrical. This concept of symmetry is useful in more general cases
and so we define :
It can be seen from the graphs below that sinh x is an odd function and cosh x is an even
function of x.
1
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
4 4
3 3
sinh x cosh x
2 2
1 1
0 0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2
-1 -1
-2 -2
-3 -3
-4 -4
Sketching the graph of a function can often be more useful than plotting it by computer.
The interesting features of a function are usually the following :
(i) At x = 0, y = −6
(ii) By inspection, we find that x = 1 is a root. Further examination then shows that the
expression factorises into,
dy 12 ± 144 −132 1
= 3 x 2 − 12 x + 11 → Turning points at x = = 2±
dx 6 3
1 1
Thus, at x = 2 − = 1.42, y = 0.38, and at x = 2 + = 2.58, y = −0.38
3 3
(iv) As x → ±∞, y → x3
2
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
We are now in a position to make a sketch of the function which actually looks like this :
0
-4 -3 -2 -1 0 1 2 3 4
-1
-2
-3
-4
-5
-6
-7
2 APPROXIMATION BY SERIES
Taylor’s theorem states that any function can be expanded as an infinite series around
the point x = a in the form,
( x − a) 2 ( x − a)3
f ( x) = f (a ) + ( x − a ) f ′(a ) + f ′′(a ) + f ′′′(a ) + ...
2! 3!
x2 x3
f (a + x) = f (a ) + x f ′(a ) + f ′′(a ) + f ′′′(a ) + ...
2! 3!
A special case is Maclaurin’s series, obtained by putting a = 0 in either of the above forms,
x2 x3
f ( x) = f (0) + x f ′(0) + f ′′(0) + f ′′′(0) + ...
2! 3!
3
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
To derive Taylor’s theorem, note that the objective is to find a power series of the form,
f (a + x) = a0 + a1 x + a2 x 2 + a3 x 3 + ...
where a0, a1, a2, a3, etc are constants to be determined. Setting x = 0 we find a0 = f(a).
Successive differentiation then gives the other coefficients :
The elementary functions of mathematics all have power series expansions which can
be derived using Taylor’s theorem and which can be found in the mathematics data book.
x2 x3 x2 x4
ex = 1 + x + + + ... cos x = 1 − + + ...
2! 3! 2! 4!
x2 x3 x4 x3 x5
ln(1 + x ) = x − + − + ... sinh x = x + + + ...
2 3 4 3! 5!
x3 x5 x2 x4
sin x = x − + − ... cosh x = 1 + + + ...
3! 5! 2! 4!
α (α −1) x 2 α (α −1)(α − 2) x 3
(1+ x)α = 1 + α x + + + ...
2! 3!
The binomial expansion is particularly useful for approximating the reciprocal of a function.
For example, if we wish to approximate the function (1 + x2)−1 when |x| << 1, we set α = −1 :
1 − 1(−1−1) ( x 2 ) 2
= 1 − (x ) + 2
+ ... = 1 − x 2 + x 4 + ...
(1+ x )
2 2!
4
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
The statement that a complicated function behaves like a simpler function in the vicinity
of a particular value of x can be made more precise by use of the O notation. For example, we
can replace the weak statement, e x ≅ 1 + x for |x| << 1, by the stronger version,
e x = 1 + x + O( x 2 )
The notation O(x2) means ‘a constant multiplied by x2 plus terms involving higher powers of
x’. Note that the ‘approximately equals sign’ has been replaced by an actual ‘equals sign’.
The value of the O notation can be seen from the series for (1 + x2)−1 given above. Thus,
1
= 1 − x 2 + O( x 4 )
(1+ x ) 2
1
= 1 − x 2 + x 4 + O( x 6 )
(1+ x ) 2
The first statement informs us that there are no terms of order x3 in the expansion as all
neglected terms are of order x4 and higher. The second statement is stronger as it informs us
that, not only are there are no terms of order x3, there are also no terms of order x5.
Change of variable
1
sin x = sin y + = sin ( y ) cos + cos ( y ) sin = (sin y + cos y )
4 4 4 2
1 y3 y2 1 y2
sin x = y− + O( y 5 ) + 1 − + O( y 4 ) = 1+ y − + O( y 3 )
2 3! 2! 2 2
5
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
To approximate f(x) for large x, try y = 1/x and examine the behaviour as y → 0.
1
Example : To approximate sin as x → ∞, we let y = 1/x. Then :
1+ x
3
1 y y 1 y
sin = sin = − + O( y5 )
1+ x 1+ y 1+ y 3! 1 + y
1 y3
sin = y [1 − y + y 2 + O( y 3 )] − [1 − 3 y + 6 y 2 + O( y 3 )] + O(y5)
1+ x 6
5 3
= y − y2 + y + O( y 4 )
6
This example shows how important it is to keep track of all contributions to the required order
of the approximation. For example, we might naively have thought that sin[y/(1+ y)] tends to
sin y as y → 0, but this is only true to the first order in y.
y2 y3
= cosh (1) 1 + + O( y 4 ) + sinh(1) y + + O( y 5 )
2! 3!
cosh(1) 2 sinh(1) 3
= cosh (1) + sinh(1) y + y + y + O( y 4 )
2 6
where,
e + e −1 e − e −1
cosh(1) = , sinh(1) = .
2 2
6
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
2
x −1 1 x x2 (−1)(−2) x x 2
(1 + e ) = 1− + + + + O( x 3 )
2 2 4 2! 2 4
1 x x2 x2
= 1− − + + O( x 3 )
2 2 4 4
1 x
= − + O( x 3 )
2 4
3 INDETERMINATE FORMS
The function φ ( x) = x tan( x) takes the value 0/0 when x = 0. This is an example of an
indeterminate form meaning that it is impossible to establish its value in the usual way by
substituting x = 0. If x is very small, even the output from a computer calculation may be
unpredictable in such a situation. We now describe two ways to overcome this difficulty.
f ( x) a0 + a 1 x + a2 x 2 + a3 x 3 + ...
φ ( x) = =
g ( x) b0 + b1 x + b2 x 2 + b3 x 3 + ...
Where a0, a1, a2 ... and b0, b1, b2 ... are the constant coefficients in the series with a0 = f(0) = 0
and b0 = g(0) = 0. Cancelling a factor x we therefore obtain,
7
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
a 1 + a2 x + a3 x 2 + ... a1
φ ( x) = → as x → 0
b1 + b2 x + b3 x 2 + ... b1
a2
φ ( x) → as x → 0
b2
and so on.
l’Hôpital’s rule is a formal version of the analysis described above. Expanding the
numerator and denominator as Taylor series about x = 0, we obtain,
f ′(0)
φ ( x) → as x → 0
g ′(0)
f ′′(0)
φ ( x) → as x → 0
g ′′(0)
and so on. l’Hôpital’s rule is useful for simple cases but direct series expansion is usually
preferable for more complicated examples.
The mathematics data book gives some useful limits of common indeterminate forms.
Particularly important are the following two cases :
ln x
(i) As x → ∞, → 0 for any real value of s > 0. Thus, ‘any power beats log’.
xs
(ii) As x → ∞, x s e − x → 0 for any real value of s. Thus, ‘exponential beats any power’.
8
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
f ′(0)
As x → 0, φ ( x) → =1
g ′(0)
x2 0 0
As x → 0, φ ( x) = → =
1 − x cotan x 1−1 0
which is no help. We therefore settle down for a long haul and substitute cotan x = cos x/sin x
because it is easier to work with sin and cos :
x 2 sin x f ( x)
φ ( x) = =
sin x − x cos x g ( x)
9
Engineering Part IA 2009-10, Paper 4, Mathematical Methods, Fast Course, J.B.Young
We now see whether the job would have been any easier using direct series expansion. Thus :
x3 x5 x5
f ( x) = x 2 sin x = x 2 x − + + O( x 7 ) = x3 − + O( x 7 )
3! 5! 6
x3 x5 x2 x4
g ( x) = sin x − x cos x = x − + + O( x 7 ) − x 1− + + O( x 6 )
3! 5! 2! 4!
x3 x5
= − + O( x 7 )
3 30
x3
As x → 0, f(x) → x + O( x ) , g(x) →
3 5
+ O( x 5 ) , and hence φ ( x) → 3
3
10