Main Index Mathematical Analysis Infinite series and products Infinite series Series with arbitrary terms
  Subject Index
comment on the page

Alternating series

These are infinite series of the form

Underoverscript[∑, n = 1, arg3] (-1)^n a _ n        & ... nbsp;       Underoverscript[∑, n = 1, arg3] (-1)^(n + 1) a _ n

with typeset structure for every typeset structure, that is the terms of the series alternate in sign.

Alternating Series Test (Leibniz):  If the alternating series typeset structure satisfies

(i)  typeset structure  for typeset structure, and

(ii) typeset structure

then the series typeset structure is convergent.

The proof follows runs along similar ideas as the proof of the convergence alternating harmonic series    with the difference that now applies the same reasoning to the expression

a _ (n + 1) - a _ (n + 2) + a _ (n + 3) - ... + (-1)^(k - 1) a _ (n + k) .

This test only tells when a series converges and not if a series will diverge. The test can be extended in such a way that it is required that eventually we will have typeset structure for all typeset structure after some point.  Thus it is possible for the first few terms of a series do not decrease.  

Example: Let typeset structure. If typeset structure, then

f^'(x) = (x^(1/2)/2 (x + 4) - x^(1/2) · 1)/(x + 4)^2 = (2/x^(1/2) - x^(1/2)/2)/(x + 4)^2 = (4 - x)/(2 x^(1/2) (x + 4)^2),

which is negative for typeset structure, and so typeset structure is decreasing for typeset structure, which implies that the series typeset structure satisfies the alternating series test when its first three terms are omitted.

[Graphics:HTMLFiles/AlternatingSeries_20.gif]

Simple examples show that condition (i) cannot be omitted.

Take a divergent positive term series whose terms go to zero, say typeset structure, and a convergent series with positive terms, say typeset structure .  Then interfuse the first with the second one to form the following series

1/2 - 1/2^2 + 1/3 - 1/3^2 + 1/4 - 1/4^2 + ...(1)

This is an alternating series which terms go to zero, but the terms do not form a monotone decreasing sequence. The resulting series diverges since the divergent harmonic series overpowers the contribution of the convergent series to force the typeset structureth partial sum off to typeset structure.

Graph of the absolute values of terms of series  (1). Red points represent the terms of series  typeset structure, blue ones the terms of typeset structure .

[Graphics:HTMLFiles/AlternatingSeries_28.gif]

The Alternating Series Error Estimate: Suppose that typeset structure is a sequence of real numbers which satisfies the hypothesis of the alternating series test. Suppose that the series typeset structureconverges to typeset structure. Let typeset structure. Then typeset structure.

A series with positive terms can be converted to an alternating series using the transformation

Underoverscript[∑, n = 1, arg3] a _ n = Underoverscript[∑, k = 1, arg3] (-1)^(n - 1) Overscript[a, ~] _ n,

where typeset structure.

To find the explicit values for alternating series is a more complicated question. For instance,

Underoverscript[∑, n = 1, arg3] (-1)^(n - 1)/n = ln 2,(2)

Graph of the absolute values of terms of series  (2). Blue points represent the absolute value of negative terms.

[Graphics:HTMLFiles/AlternatingSeries_38.gif]

The series on the left hand side of  (2) is sometimes called the alternating harmonic series.

Euler discovered the following formula for typeset structure :

Underoverscript[∑, n = 1, arg3] (-1)^(n - 1)/(2 n - 1) = 1 - 1/3 + 1/5 - 1/7 + ... = π/4(3)

[Graphics:HTMLFiles/AlternatingSeries_42.gif]

We also have

Underoverscript[∑, n = 0, arg3] (-1)^n/n ! = e^(-1),(4)
Underoverscript[∑, n = 1, arg3] (-1)^(n - 1)/n^2 = π^2/12 .(5)

B. Schmuland  [1]   studied properties of the random harmonic series

Underoverscript[∑, n = 1, arg3] ε _ n/n(6)

where the typeset structure, typeset structure, are independent random variables taking the values  typeset structure and typeset structurewith common distribution typeset structure. He noticed that Kolmogorov’s three series theorem  or martingale convergence theorem imply that the sequence of typeset structureth partial sums of (6) converges almost surely. He proved that the sum (6) has a uniform distribution on typeset structure and describes some interesting properties of the convergent as random variable.

References

[1]  Schmuland, B. (2003). Random harmonic series. American Mathematical Monthly, 110, 407-416.

Cite this web-page as:

Štefan Porubský: Alternating series.

Page created  .