ON REARRANGEMENTS OF SERIES H. M. SENGUPTA

Similar documents
MATH115. Infinite Series. Paolo Lorenzo Bautista. July 17, De La Salle University. PLBautista (DLSU) MATH115 July 17, / 43

ON THE DENSITY OF SOME SEQUENCES OF INTEGERS P. ERDOS

THE FIXED POINT PROPERTY FOR CONTINUA APPROXIMATED FROM WITHIN BY PEANO CONTINUA WITH THIS PROPERTY

Absolute Convergence and the Ratio & Root Tests

Math 143 Flash Cards. Divergence of a sequence {a n } {a n } diverges to. Sandwich Theorem for Sequences. Continuous Function Theorem for Sequences

("-1/' .. f/ L) I LOCAL BOUNDEDNESS OF NONLINEAR, MONOTONE OPERA TORS. R. T. Rockafellar. MICHIGAN MATHEMATICAL vol. 16 (1969) pp.

Week 2: Sequences and Series

Copyright c 2007 Jason Underdown Some rights reserved. statement. sentential connectives. negation. conjunction. disjunction

GENERATING FUNCTIONS OF CENTRAL VALUES IN GENERALIZED PASCAL TRIANGLES. CLAUDIA SMITH and VERNER E. HOGGATT, JR.

AxiOM F. If P is a point of a region R and X is a point distinct from P. THEOREM 2. If, in a space satisfying Axioms 0, 1 and 2, the boundary

Analysis II: Basic knowledge of real analysis: Part IV, Series

A Geometrical Proof of a Theorem of Hurwitz.

T(R) = R. THEOREM 3.4: If T(M) C M is a continuous transformation of a compact

n»i \ v f(x + h)-f(x-h)

MAT 570 REAL ANALYSIS LECTURE NOTES. Contents. 1. Sets Functions Countability Axiom of choice Equivalence relations 9

SEPARABILITY OF STOCHASTIC PROCESSES

ON THE SPACE OF INTEGRAL FUNCTIONS. IV

Real Analysis Math 131AH Rudin, Chapter #1. Dominique Abdi

A CHARACTERIZATION OF INNER PRODUCT SPACES

INTRODUCTION TO REAL ANALYSIS II MATH 4332 BLECHER NOTES

Definition 2.1. A metric (or distance function) defined on a non-empty set X is a function d: X X R that satisfies: For all x, y, and z in X :

MATH 117 LECTURE NOTES

ON CONVERGENCE OF STOCHASTIC PROCESSES

ON CONVERSE GAP THEOREMS

MATH 409 Advanced Calculus I Lecture 7: Monotone sequences. The Bolzano-Weierstrass theorem.

ON THE OSCILLATION OF THE DERIVATIVES OF A PERIODIC FUNCTION

^-,({«}))=ic^if<iic/n^ir=iic/, where e(n) is the sequence defined by e(n\m) = 8^, (the Kronecker delta).

ON THE NUMERICAL RANGE OF AN OPERATOR

IRREGULARITIES OF DISTRIBUTION. V WOLFGANG M. SCHMIDT

2 8 P. ERDŐS, A. MEIR, V. T. SÓS AND P. TURÁN [March It follows thus from (2.4) that (2.5) Ak(F, G) =def 11m 1,,,,(F, G) exists for every fixed k. By

Iowa State University. Instructor: Alex Roitershtein Summer Homework #5. Solutions

GAUSSIAN PROCESSES GROWTH RATE OF CERTAIN. successfully employed in dealing with certain Gaussian processes not possessing

CONTINUA AND VARIOUS TYPES OF HOMOGENEITY(1)

Chapter 4. Measure Theory. 1. Measure Spaces

ON THE "BANG-BANG" CONTROL PROBLEM*

Polynomial expansions in the Borel region

THE RANGE OF A VECTOR-VALUED MEASURE

2 K cos nkx + K si" nkx) (1.1)

Chapter 10. Infinite Sequences and Series

THE PRESERVATION OF CONVERGENCE OF MEASURABLE FUNCTIONS UNDER COMPOSITION

MINIMAL.4-SETS, INFINITE ORBITS, AND FIXED ELEMENTS

From Calculus II: An infinite series is an expression of the form

Czechoslovak Mathematical Journal

(4) cn = f f(x)ux)dx (n = 0, 1, 2, ).

P-adic Functions - Part 1

lim f f(kx)dp(x) = cmf

MATHEMATICS AND GAMES* Nagayoshi lwahori University of Tokyo

AP Calculus Chapter 9: Infinite Series

THE G-FUNCTIONS AS UNSYMMETRICAL FOURIER KERNELS. II

ON THE MAZUR-ULAM THEOREM AND THE ALEKSANDROV PROBLEM FOR UNIT DISTANCE PRESERVING MAPPINGS

FUNCTIONS THAT PRESERVE THE UNIFORM DISTRIBUTION OF SEQUENCES

(i) tn = f î {nk)tk(i - ty-\dx{t) (»>o).

2. Linear recursions, different cases We discern amongst 5 different cases with respect to the behaviour of the series C(x) (see (. 3)). Let R be the

AN OCCUPATION TIME THEOREM FOR A CLASS OF STOCHASTIC PROCESSES^)

DEFINITION IN TERMS OF ORDER ALONE IN THE LINEAR CONTINUUM AND IN WELL-ORDERED SETS*

On completing partial Latin squares with two filled rows and at least two filled columns

ON TOPOLOGISING THE FIELD C(t)

Math 172 HW 1 Solutions

COMPOSITION OF LINEAR FRACTIONAL TRANSFORMATIONS IN TERMS OF TAIL SEQUENCES

NATIONAL BOARD FOR HIGHER MATHEMATICS. M. A. and M.Sc. Scholarship Test. September 17, Time Allowed: 150 Minutes Maximum Marks: 30

A CONVEXITY CONDITION IN BANACH SPACES AND THE STRONG LAW OF LARGE NUMBERS1

Analysis Finite and Infinite Sets The Real Numbers The Cantor Set

A NOTE ON CONTRACTIVE MAPPINGS

HW 4 SOLUTIONS. , x + x x 1 ) 2

NATIONAL BOARD FOR HIGHER MATHEMATICS. Research Scholarships Screening Test. Saturday, January 20, Time Allowed: 150 Minutes Maximum Marks: 40

Contents Ordered Fields... 2 Ordered sets and fields... 2 Construction of the Reals 1: Dedekind Cuts... 2 Metric Spaces... 3

ON AN INFINITE SET OF NON-LINEAR DIFFERENTIAL EQUATIONS By J. B. McLEOD (Oxford)

MORE ON CONTINUOUS FUNCTIONS AND SETS

Oscillation Criteria for Delay Neutral Difference Equations with Positive and Negative Coefficients. Contents

ON THE VOLUME OF THE INTERSECTION OF TWO L"p BALLS

SOME APPLICATIONS H. F. BLICHFELDT

Problems and Solutions

A NOTE ON FUNCTION SPACES GENERATED BY RADEMACHER SERIES

QUOTIENTS OF F-SPACES

ASYMPTOTIC VALUE OF MIXED GAMES*

Mini-Course on Limits and Sequences. Peter Kwadwo Asante. B.S., Kwame Nkrumah University of Science and Technology, Ghana, 2014 A REPORT

BOUNDS ON THE EXPECTATION OF FUNCTIONS OF MARTINGALES AND SUMS OF POSITIVE RV'S IN TERMS OF NORMS OF SUMS OF INDEPENDENT RANDOM VARIABLES

BOUNDEDNESS OF SOLUTIONS TO FUNCTIONAL INTEGRO-DIFFERENTIAL EQUATIONS ZUOSHENG HU. (Communicated by Kenneth R. Meyer)

ON THE ORLICZ-PETTIS PROPERTY IN NONLOCALLY CONVEX F-SPACES

ON DIVISION ALGEBRAS*

ANNALES DE L INSTITUT FOURIER

THE EXTENSION OF AN ARBITRARY BOOLEAN ALGEBRA TO AN IMPLICATIVE BOOLEAN ALGEBRA

ON DIFFERENCE SETS WHOSE PARAMETERS SATISFY A CERTAIN RELATION

Some analysis problems 1. x x 2 +yn2, y > 0. g(y) := lim

Kotoro Tanahashi and Atsushi Uchiyama

TWO KINDS OF CHAOS AND RELATIONS BETWEEN THEM. 1. Introduction

Problems for Chapter 3.

(1) H* - y\\ < (1 + r)(x - y) - rt(tx - ra)

1. If 1, ω, ω 2, -----, ω 9 are the 10 th roots of unity, then (1 + ω) (1 + ω 2 ) (1 + ω 9 ) is A) 1 B) 1 C) 10 D) 0

Math 430/530: Advanced Calculus I Chapter 1: Sequences

LECTURE 10: REVIEW OF POWER SERIES. 1. Motivation

0 < f(y) - fix) - (x*, y - x) < e\\y - x\\

General Notation. Exercises and Problems

STRONGLY EXTREME POINTS IN LPfa X)

Copyright 1963, by the author(s). All rights reserved.

PROBABILISTIC TURING MACHINES AND COMPUTABILITY

A SINGULAR FUNCTIONAL

Analysis Qualifying Exam

Appendix B for The Evolution of Strategic Sophistication (Intended for Online Publication)

Two-Step Iteration Scheme for Nonexpansive Mappings in Banach Space

Transcription:

O REARRAGEMETS OF SERIES H. M. SEGUPTA In an interesting paper published in the Bulletin of the American Mathematical Society, R. P. Agnew1 considers some questions on rearrangement of conditionally convergent series. He considers the metric space E in which a point x is a permutation Xi, *», ) of the positive integers and in which the distance p(x, y) between two points **= x2, #«, ) and y= (yi, y2, y%, ) of E is given by the Fr6chet formula " 1 1 xn - y I Agnew proves that the space is of the second category at each of its points. He also considers the following problem: Let C\ + d + Cs-\- - be a conditionally convergent series of real terms. Denote C by C(«). To each xg.e there corresponds a rearrangement 2~ln-i C(xn) of the series C(n) or C and also to each rearrangement of the series corresponds a point i. Thus the ways in which the series may be rearranged form a set which has the potency of E. It is well known that x(e.e exists for which C(xn) converges to any preassigned number, diverges to +» or to or oscillates with prescribed upper and lower limits. The set A of xe:e for which C(xn) converges is therefore a proper subset of E. Professor Agnew considers the nature of the set A and proves that it is of the first category so that the complementary set E A is of the second category. In point of fact he proves more than that. He shows that the set of points xce for which 2~ln-i C(x») has unilaterally bounded partial sums is of the first category. His first theorem runs as follows: For each except those belonging to a set of the first category, lim inf C(xn) = co, lim sup C(xn) = + o. JV-»«o n_l JT-»» n-1 It is possible to add something more to the above in regard to the nature of the set A. In fact we can prove the following theorem: Received by the editors ovember 28,1947 and, in revised form, August 16,1948. 1 Ralph Palmer Agnew, On rearrangement of series, Bull. Amer. Math. Soc. vol. 46 (1940) pp. 797-799. 71

72 h. m. sengupta [February Theorem 1. For each x E, except those of a set 73 (Fr in E) that is an outeimiting set of sets closed in E, we have if lim inf C(xn) = oo, lim sup C(z ) = 4-». JV->» n_l ->*> n-1 Combining the above with the result given in Agnew's paper, we may assert that the set of points x for which If it lirn inf C(xn) = oo and lim sup C(xn) = 4- oo -*<o n 1 ->*> _1 is the inner limiting set of a sequence of everywhere dense sets open in E. Proof. We require the following ideas for the proof of our theorem. Consider the series T^T-i C(xn). Let Sj, s2, Si,, Sir, " *. where sn= fli C(xn), be the partial sums. Let us denote the maximum of the numbers (si, s»,, sif) by Uif. Then si, i/2, E/», is a monotone nondecreasing sequence of real numbers, and lim^-.«, Un exists either as a finite number or as + oo. We write lim^..., Un as lim.y_0o upper bound Jli C(xn). In a similar manner if denotes min (si, 52,, ss), then S%, «2, Ma, is a monotone nonincreasing sequence of real numbers and limat.«, un exists either as a finite number or as oo. We denote limjv^oo Un by the symbol limjr..«, lower bound Yn-i (*«) ow, let A>0 be large at pleasure. We define 73» to be the set of points of E for which limar..«, upper bound 2~l»-i C(xn) h. ow let i, 2, be a convergent sequence of points of E all belonging to 73. Let their limit point also belong to E. We shall prove that also belongs to 73*. Let Zn=-(x?\ 4n\ xf, ) and =(xu x2, xs, ). The number complexes (ar?0, x??, x3n), ) for n 1, 2, 3, and also the complex (xi, x2, xit ) are different permutations of the positive integers. If now does not belong to 73* we must have lim upper bound C(xr) > h. -f" n-1 So, for sufficiently large 7Y, say =M' where M' is some positive integer depending on and h, we must have Un>h. In particular, therefore, Uw>h. But Uw is the maximum of (si, ft, 5»,, iaf). so there is a positive integer M (1 = M^M') such that SM>h, that is, there is a positive integer J17 for which Sm = C(*0 + C(*2) + + C(xm) > h. if

O REARRAGEMETS OF SERIES 73 We may now take a sphere of sufficiently small radius to ensure that the first M elements of all points of E that lie in the interior of the sphere are identical with those of in value and in order. So for all these points the corresponding rearranged series are'such that the Mth partial sum of each of them equals Sm = C(xi) + C(x2) 4- + C(xm) and so exceeds h in value. ow, since lim =, for all sufficiently large n, say ti P where P is a suitable positive integer, lies within the sphere referred to above. So for»=p, P+l, C(xT) = C(*i) + C(x2) + + C(xM) > k P4-2,. Therefore tr,. lim upper bound C(x" ) > h (» = P, P + 1,.. ), -"> that is, p, i-p+i, lp+2, all belong to E Bh which is contrary to the hypothesis. So belongs to 5», that is, the set 75 is closed in E. In a like manner, we may prove that if B-h stands for the set of points of E for which lim lower bound C( xr) St h ->» and if {fi, f»,», } is a convergent sequence of points of E all belonging to 73_* with f as the limit which itself belongs to E, the f G-B-». So, is closed in. It follows therefore that for any positive integer n, the set 73 +7i_B is closed in E, that is, the set of points of E, for which or tr lim upper bound C( xr) g» -ko lim lower bound C(*v) =» W-«> is closed in E. The set 5 referred to above is, therefore, given by B = 00 n-1 (7i + B-n). So, 73 is a set F. in, that is, the outer limiting set of a sequence of sets closed in E.

74 h. m. sengupta [February It follows that the set of points for which lim inf C(xn) = and lim sup 2~l'C(xn) = 4-00 n-l n-l is Gt in E. We next propose to derive some simple properties about the power of E. Theorem 2. The set E has the power c of the continuum. Proof. Let p be the power of the set E. Let us associate the point x = (xi, Xi, Xi, ) of the set E to the irrational number a which is the value of the simple continued fraction 1 1 1 Xl + x2 + xt +. Thus to each point x(e.e there corresponds an irrational number a in 0<a<l and to two distinct points x and y of E correspond two distinct irrational numbers in the above range. Thus the set E is equivalent to a proper subset of the real numbers. Therefore p C. On the other hand, to any real number ß (0<ß<l) there corresponds a rearrangement of the conditionally convergent series T«i C such that the rearranged series converges to ß. In fact there are an infinity of such rearrangements with the sum ß. So to any real number correspond an infinity of distinct points of E. Again to any other real number y, there correspond an infinity of other distinct points of E, each of which is different from any of those corresponding to ß. Thus the set of real numbers is equivalent to a proper subset of the aggregate of points of E. Therefore p=c. Combining these two, we have p=c. Theorem 3. Every point of the metric space E is of degree c (power of the continuum) in E* Proof. Take any point x = (*i, x2, xit ) of E. Then C(xi) + C(xt) + C(*,) + ; is a rearrangement of the given conditionally convergent series. Let 7Y be a positive integer. Let 0<a<l and let 2 The proof of Theorem 3 is due to the referee. It may be remarked that a proof of the above modelled on the plan of proof of Theorem 2 may also be easily given.

i95o] on rearrangements of series 75 a =.OLiaictzcti be the dyadic expansion of a which may terminate with zeros but not with ones. Let *ia) =xn («= 1, 2, 3,, 7Y+1) and if ap=0, let and if ap = 1, let («0 («) Xn+2p = Xff+tp, Xtf+ip+l = Xn+tp+1 <«> («) %+tp X+ip+u Xtr+tp+i = xtr+ip- Then to different numbers a correspond different points (a) (a) (a) (a) * = (*i, x»,*$, ). All points x(a) lie within the sphere with center * and radius And since lim C(xn) = 0, the series 2~s. C(x[")) + C(.xiia))+C(x(t")) + -" has, for each or, exactly the same limits of oscillation as the series c(*»). This proves that the points of E that lie within the sphere S(x, 2~) has the power of the continuum. It proves more. It easily leads to the theorem: Theorem 4. Every point of E is the limit point of a set of points of E of power c at which the rearrangements of the conditionally convergent series behave in any prescribed manner. Proof. Take any point x = (xu x2, x3, ). Let / and L be any two real numbers,» and 4-00 not excepted, such that <x> gl=l ^ 4-00 In a sphere with x as center and 2~s as radius there exists a point X = (aci, X2,, xif+u Xff+i, X+t, ) at which the rearranged series has / and L as its limits of oscillation. Again, by the above, in a sufficiently small sphere with X as center lying entirely within the first sphere, an infinite set of points of the power C exist at which the rearranged series has the same limits of oscillation / and L. Thus in every neighbourhood of x, there exists a set of points of the power of the continuum, at which the rearranged series has any prescribed limits of oscillation. Dacca, India