The Structure of Real Sequences and Series
[44/100%]Let the terms of the sequence satisfy the condition
then the sequence
either converges to its lower bound or diverges properly to .
For a special case see M. Fekete: Math. Z. Vol. 17, p. 233 (1939).
Assume that the terms of the sequence satisfy the condition
Then
exists; is finite and we have
If the general term of a series which is neither convergent nor properly divergent tends to the partial sums are everywhere dense between their lowest and their highest limit points.
L. Fejér: C. R. Acad. Sci. (Paris) Sér. A-B, Vol. 142, pp. 501-503 (1906).
Let , and the series be divergent. Put and denote by the largest integer . Find the limit points of the sequence
Cf.G. Szegő, Problem: Arch. Math. Phys. Ser. 3, Vol. 23, p. 361 (1914). Solved by P. Veress: Arch. Math. Phys. Ser. 3, Vol. 25, p. 88 (1917).
Assume that there exists for the sequence a sequence of positive numbers , converging to , for which
Then the numbers are everywhere dense between their lowest and highest limit points.
G. Pólya: Rend. Circ. Mat. Palermo Vol. 34, pp. 108-109 (1912).
Let be positive integers, . The set of limit points of the sequence
consists of a closed interval (of length 0 if the limit exists).
Problem 102.
A subsequence whose terms are the successive partial sums of an absolutely convergent series can be picked out from every convergent sequence.
A sequence that diverges to contains a minimum (i.e.there exists a such that for all ).
A convergent sequence has either a maximum or a minimum or both.
Let be a sequence of positive numbers (positive in the sense of ) and let . Then there are infinitely many subscripts for which is smaller than all the terms preceding , (, ).
Let be a sequence of positive numbers (positive in the sense of ) and let . Then there are infinitely many subscripts for which is larger than all the terms following , (, ). (Not only the conclusion but also the hypothesis is different from the one in problem 107.)
Given two sequences
satisfying the conditions
Then there are infinitely many subscripts such that two different kinds of inequalities hold simultaneously:
G. Pólya: Math. Ann. Vol. 88, pp. 170-171 (1923).
If the sequence tends to and if is larger than its minimum [problem 105] then there exists a subscript (or several subscripts ), , so that the quotients
are and the infinitely many quotients
are all .
The quantities in question can be interpreted as the slopes of certain connecting lines between the points with cartesian coordinates ; . This interpretation leads to a geometric proof of the statement. [Concerning problems 110-112 cf.A. Wiman: Acta Math. Vol. 37, pp. 305-326 (1914); G. Pólya: Acta Math. Vol. 40, pp. 311-319 (1916); G. Valiron: Ann. Sci. Ecole Norm. sup. (Paris) Ser. 3, Vol. 37, pp. 221-225 (1920); W. Saxer: Math. Z. Vol. 17, pp. 206-227 (1923).]
Assume that the sequence satisfies the sole condition
Let be larger than (). Then there exists a subscript , , such that all these inequalities hold simultaneously:
If tends to infinity then so does .
Cf.G. Pólya, Problem: Arch. Math. Phys. Ser. 3, Vol. 24, p. 282 (1916).
Let the sequence satisfy the two conditions
and assume . Then there exists a subscript , , such that the inequalities
hold simultaneously. If tends to , then tends to infinity.
Show that
Let be arbitrary real numbers, . If there exists a positive distance such that , , , then the convergence exponent of is at most .
Let be larger than the convergence exponent of the sequence . Then there exist infinitely many subscripts for which the inequalities
are satisfied.
See problem 107.
Assume that the convergence exponent of the sequence is positive and . Then there are infinitely many subscripts such that the two types of inequalities
hold simultaneously.
See problem 109.
Suppose . For what values of , , is the -th term of the series
larger than all the other terms, ?
Assume
Then arbitrarily large values of and can be found for which the following inequalities hold simultaneously [problem 111]:
At the root of this fact, and of problem 122, lies the comparison of two power series
For an everywhere convergent power series in which is not merely a polynomial the central subscript tends to with .
The subscript of the maximum term increases as increases.
One might consider this situation as somewhat unusual: in the course of successive changes the position of maximum importance is held by more and more capable individuals.
The series
with positive coefficients and finite radius of convergence (, ) is such that one term after the other, all terms in turn, become maximum term. Then is the radius of convergence of the series
The dominance of the maximum term is more pronounced in an always convergent series than in one that does not always converge (it is strongest in a polynomial). More exactly: let the radius of convergence of the power series
be infinite and that of the power series
be finite. Suppose , , The coefficients be such that all the terms become in turn maximum term [problem 120]. Then a value can be determined for certain arbitrarily large positive such that for these corresponding values the respective series have the same central subscript. Let the common central subscript be . Then all the following inequalities hold simultaneously:
Consider the maximum term of . [l.c.solution 110.]
If there are values to which no corresponds in the sense of problem 122 then they are ``rare''. They have a finite logarithmic measure, i.e.the set of points , exceptional value, may be covered by countably many intervals of finite total length.
l.c.solution 110.
From the harmonic series
remove all terms that contain the digit 9 in the decimal representation of the denominator. The resulting subseries is convergent.
A. J. Kempner: Amer. Math. Monthly Vol. 21, pp. 48-50.
If all the subseries of a series converge then the series is absolutely convergent.
Let and denote positive integers. Must the convergent series be absolutely convergent if all its subseries of the form
(subscripts in arithmetic progression) converge?
K. Knopp: J. reine angew. Math. Vol. 142, pp. 292-293 (1913).
Let and be integers, , . Must the convergent series be absolutely convergent if all its subseries of the form
(subscripts in geometric progression) are convergent?
Let denote a polynomial assuming integral values for integral , Assume that the degree is and that the coefficient of is positive (). The values form a generalized arithmetic progression of order ; since only a finite number of terms of the progression can be negative. Must a convergent series converge absolutely if all its subseries whose subscripts form a generalized arithmetic progression
(omitting the terms with negative subscripts) converge?
A polynomial assuming integral values for every integer need not itself have integral coefficients -- e.g. is such a polynomial. This is the class of polynomials meant here (cf.Part VIII, Chapter 2).
If the series converges absolutely and if every subseries
has the sum then .
A. Haar.
Consider the set of points determined by all the subseries of
This set is perfect and nowhere dense (closed and dense in itself, but nowhere dense in the set of all real numbers).
We have to consider all the subseries, finite and infinite, including the ``empty'' subseries to which we attribute the sum 0. G. Cantor, cf.E. Hewitt and K. Stromberg: Real and Abstract Analysis. Springer: New York 1965, pp. 70-71.
Let the terms of the convergent series
satisfy the inequalities
Then it is possible to represent any number in the half-closed interval by an infinite subseries:
Cf.S. Kakeya: Proc. Tokyo math.-phys. Soc. Ser. 2, Vol. 7, p. 250 (1914); Tôhoku Sc. Rep. Vol. 3, p. 159 (1915).
Find the series that satisfies the conditions
and verify that in this case every mentioned in problem 131 can be represented by one infinite subseries only.
By rearranging the factors of the infinite product
we obtain the infinite product
in which blocks of factors greater than 1 alternate with blocks of factors smaller than 1. (Factors of the same kind remain in the ``natural'' order.) Show that
G. Pólya, Problem: Amer. Math. Monthly Vol. 51, pp. 533-534 (1944). Solution Amer. Math. Monthly Vol. 53, pp. 279-282 (1946).
By rearranging the factors of the infinite product
we obtain the infinite product
in which blocks of factors greater than 1 alternate with blocks of factors smaller than 1. (Factors of the same kind remain in the ``natural'' order.) Show that
G. Pólya, Problem: Amer. Math. Monthly Vol. 51, pp. 533-534 (1944). Solution Amer. Math. Monthly Vol. 53, pp. 279-282 (1946).
By rearranging the terms of the infinite series
we obtain the infinite series
in which blocks of positive terms alternate with blocks of negative terms. (Terms of the same kind remain in the ``natural'' order, steadily decreasing in absolute value.) Show that
See problem 132.1.
By rearranging the terms of the infinite series
we obtain the infinite series
in which blocks of positive terms alternate with blocks of negative terms. (Terms of the same kind remain in the ``natural'' order, steadily decreasing in absolute value.) Show that
See problem 132.1.
If one of the two to each other complementary subseries of a convergent series is convergent the other is convergent too. A rearrangement which only shifts the two subseries relatively to each other does not change the sum of such a series.
If one of the two to each other complementary subseries of a conditionally convergent series diverges to then the other diverges to . Provided that all the terms of one of these two subseries are of the same sign it is possible to obtain an arbitrary sum for the whole series by shifting the two subseries relatively to each other.
Knopp, pp. 318; cf.also W. Threlfall: Math. Z. Vol. 24, pp. 212-214 (1926).
It is not possible to accelerate by rearrangement the divergence of a divergent series with positive monotone decreasing terms.
By rearranging the series we can slow down arbitrarily the divergence of a divergent series with positive terms which tend to . More explicitly: assume
Then there exists a rearranged series such that for
Assume that
and is divergent. Let . By a rearrangement that leaves all the negative terms at their places the sum can be realized; by a rearrangement that leaves all the positive terms at their places the sum can be realized [problem 136].
W. Sierpiński: Bull. int. Acad. pol. Sci. Lett. Cracovie 1911, p. 149.
Assume that , and that the series
is divergent and the series
where is or , is convergent. If under these conditions a certain percentage of terms is positive then it is . More precisely:
E. Cesàro: Atti Accad. Naz. Lincei Rend. Cl. Sci. Fis. Mat. Natur Ser. 4, Vol. 4, 2nd Sem. p. 133 (1888); J. Bagnera: Darboux Bull. Ser. 2, Vol. 12, p. 227 (1888). Cf.also G. H. Hardy: Mess. Math. Ser. 2, Vol. 41, p. 17 (1911); H. Rademacher: Math. Z. Vol. 11, pp. 276-288 (1921).
Suppose , and that the series
where is or , is convergent. Then
Notice the two well-known extreme cases and . E. Lasker.