Sei sulla pagina 1di 5

SHES 2303 MATHEMATICS IN BIOLOGY

1 Limits of Sequences of Numbers

Let us consider a sequence of n numbers

{an } = {a1 , a2 , a3 , · · · , an }.

For example, if an = n, then {an } is just the sequence of positive integers from 1 to n.
We are interested in the behaviour of an when n becomes large. Here are some examples:
an = n becomes arbitrarily large as n tends to infinity; the sequence {an } is therefore
divergent. On the other hand, if an = 1/n2 , which tends to 0 (a constant) as n tends to
infinity, then the sequence {an } is convergent.

EXAMPLE 1. Is the sequence


 
1 1 1 n+1 1
1, − , , − , · · · , (−1) ,
2 3 4 n

convergent or divergent?

SOLUTION. As n gets larger, the fractional part gets smaller, with its sign oscillating
between +1 and -1. It manages to converge to 0 as n tends to infinity. ♣

Knowing the convergence or divergence of a sequence of numbers is important for


finding infinite sums. If a sequence is divergent, then the sum of all the elements in this
divergent sequence must also be infinite. On the other hand, if a sequence is convergent,
the sum of all the elements in this convergent sequence may or may not be finite. The
infinite sum of {an } is generally called an infinite series. It is defined as the limit of the
sum of {an } (the partial sum). For example, suppose an = (1/2)n−1 . The sum of all n
terms in {an } is therefore
n  i−1
X 1 1 1 1
Sn = =1+ + 2 + · · · n−1 .
i=1
2 2 2 2

The limit of Sn as n tends to infinity is written as


∞  i−1
X 1
lim Sn = ,
n→∞
i=1
2

with 2 as the answer (geometric series).

1
The geometric series warrants special attention because it appears very often in math-
ematical work. In a general, we write a geometric series as

X
2 n−1
a + ar + ar + · · · + ar + ··· = arn−1 ,
n=1

where a and r are fixed real numbers, with a 6= 0 and r can either be positive or negative.
For |r| ≥ 1, this series diverges, since the nth term increases as n increases. For |r| < 1,
the series converges to a/(1 − r). The proof is left as an exercise.

2 Determining Convergence and Divergence of Infi-


nite Series

When we have an infinite series, we are usually concerned with two issues: does it
converge? and if it converges, what is its sum? A test for divergence uses the simple fact
that if a series converges, limn→∞ an must
P be 0.2 So if the latter is not 0 or does not exist,
then the series diverges. For example, ∞ n=1 n diverges because n 2
→ ∞. Also,

X −n
,
n=1
n + 1

diverges, because limn→∞ −n/(n + 1) = −1 6= 0. Another example is



X
(−1)n+1 ,
n=1

which diverges because limn→∞ (−1)n+1 does not exist.


While we can prove divergence by showing that the limit of the nth term as n → ∞
does not exist, or is nonzero, we cannot prove convergence by simply noting that the
limit of the nth term is 0. The latter is a necessary, but not sufficient condition for
convergence. To check convergence, we can use several tests, not all of which will be
covered here.

TESTS FOR CONVERGENCE OF INFINITE SERIES

1. The Integral Test: It can be shown that



X Z ∞
an ≤ C + f (x)dx,
n=N N

where an = f (n), C is some constant, and f is a continuous, positive, decreasing function


of x for all x ≥ N . Because of this, the series and the integral converge or diverge together.
2
EXAMPLE 3. The sum ∞
X 1
,
n=1
n2
converges because
∞ Z ∞
X 1 1
2
<1+ dx.
n=1
n 1 x2
(sketch a graph to show this) The value of the integral is 1, so we know that the sum on
the left-hand side must be less than 2 (i.e. it converges to some value less than 2). It can
be shown using advanced calculus that the exact value is π 2 /6 ≈ 1.644. ♣

EXAMPLE 4. The sum ∞


X 1
,
n=1
n
diverges because
∞ Z ∞
X 1 1
<1+ dx.
n=1
n 1 x
Since the integral is ∞, the series diverges. ♣

2. The Direct Comparison


P Test: P
a) If an ≤ cn and Pcn converges, thenP an converges.
b) If an ≥ dn and dn diverges, then an diverges.

EXAMPLE 5. Let an = 1/n!. If we choose cn = 1/2n , then an ≤ cn (verify). Since



X 1
= 1,
n=1
2n

The series ∞
X 1
,
n=1
n!
converges, and must be less than 1 (the true value is e − 1 as we shall see later). ♣

3
EXAMPLE 6. Check if the following series converges or not:

X 3
√ .
n=1
n+ n


SOLUTION. Write an = 3/(n + n), and choose dn = 3/(n + n) = 3/(2n). Clearly,
dn > an . Now,

X 3
,
n=1
2n
diverges (see Example 4). This implies that

X 3
√ ,
n=1
n+ n

must also diverge. ♣

3 Taylor & Maclaurin Series

Have you ever wondered how values of like e and log 0.5 are computed? It turns out that
there is a nice theory that allows us to compute such values to any degree of accuracy
that we need.

DEFINITION 1 (TC). Let f be a differentiable function with derivatives of all orders


over some interval with a as an interior point. We call the following series

X f (k) (a) f 00 (a) f (n) (a)
(x − a)k = f (a) + f 0 (a)(x − a) + (x − a)2 + · · · + (x − a)n + · · · ,
k=0
k! 2! n!

the Taylor expansion of f (x) about a. The Maclaurin series generated by f is just the
Taylor series with a = 0.

EXAMPLE 7. Suppose we have f (x) = x. To find the Taylor series generated by f
about a = 1, we need to find the nth derivative of f . Now, we know that
1
f 0 (x) = √
2 x
1
f 00 (x) = −
2 · 2x3/2
3
f 000 (x) =
2 · 2 · 2x5/2
3·5
f (4) (x) = −
2 · 2 · 2 · 2x5/2
4
By induction, we know that

(2n − 3)!!
f (n) (x) = (−1)n+1 ,
2n x(2n−3)/2
for n = 2, 3 . . ., where (2n − 3)!! = (2n − 3)(2n − 5) · · · 3 · 1. We can therefore write the
required Taylor series as

√ 1 X (2k − 3)!!
x = 1 + (x − 1) + (−1)k+1 k (2k−3)/2 (x − 1)k .
2 k=2
2 k!x

Suppose we wish to find 1.01. You have already learned how to do this using linearisa-
tion. Actually,
√ the latter is a special case of using only the first two terms of the Taylor
series of x to approximate the true value. The higher order terms can be ignored √ be-
k k
cause (x − 1) = 0.01 ≈ 0. The so-called first order Taylor approximation of 1.01 is
therefore 1.005. The true value is 1.004988 (to 6 decimal places). ♣

In many cases, it is sufficient to just use the Maclaurin series. Below is a list of
Maclaurin series for very commonly encountered functions, from which several other
more series can be derived. You should try to confirm their correctness using Taylor
expansion.


1 X
= 1 + x + x2 + · · · + xn + · · · = xn ; |x| < 1
1−x n=0

x2 xn X xn
ex = 1 + x + + ··· + + ··· = ;
2! n! n=0
n!

x3 x5 x2n+1 X x2n+1
sin x = x − + − · · · + (−1)n + ··· = (−1)n (−1)n .
3! 5! (2n + 1)! n=0
(2n + 1)!

From the first series, if we replace x with −x, then we have



1 X
= 1 − x + x2 − · · · + (−1)n xn + · · · = (−1)n xn ;
1+x n=0

if we integrate both sides with respect to x, we obtain



x2 x3 n−1 x
n
n x
n+1 X
n−1 x
n
log(1 + x) = x − + − · · · + (−1) + (−1) + ··· = (−1) ,
2 3 n n+1 n=0
n

which is valid for −1 < x ≤ 1. We can construct Maclaurin series for other functions by
appropriately applying integration or differentiation operation on the three basic expan-
sions given above. These are given as exercises in the tutorial.

Potrebbero piacerti anche