# Absolute convergence The answer is that because S is not absolutely convergent, rearranging its terms changes the value of the sum. This means {style d'affichage S_{1}} et {style d'affichage S_{2}} are not equal. En réalité, the series {style d'affichage 1-1+1-1+...} does not converge, so S does not have a value to find in the first place. A series that is absolutely convergent does not have this problem: rearranging its terms does not change the value of the sum.

Explanation This is an example of a mathematical sleight of hand. If the terms of S are rearranged in such a way that every term remains in its original position, one finds that S is either the infinite series {displaystyle S=1-1+1-1+...+1-1+1-1} or with equal possibility, ce {displaystyle S=1-1+1-1+...+1-1+1} Evaluating S as before, by grouping every -1 with the +1 preceding it or by grouping every +1 except the first with the -1 preceding it, gives in the first case: {style d'affichage S_{1}=(1-1)+....+(1-1)=0+....+0=0} {style d'affichage S_{2}=1+(-1+1)+....+(-1+1)-1=1+0+....+0-1=1-1=0} et dans le second cas: {style d'affichage S_{1}=(1-1)+....+(1-1)+1=0+....+0+1=1} {style d'affichage S_{2}=1+(-1+1)+....+(-1+1)=1+0+...+0=1} This reveals the trick: the definition of S was interpreted as defining its last term as negative when evaluating {style d'affichage S_{1}=0} but positive when evaluating {style d'affichage S_{2}=1} when in fact the definition of S didn't define (and the rearrangement was independent of) either option.

Definition for real and complex numbers A sum of real numbers or complex numbers {textstyle sum _{n=0}^{infime }un_{n}} is absolutely convergent if the sum of the absolute values of the terms {textstyle sum _{n=0}^{infime }|un_{n}|} converge.

Sums of more general elements The same definition can be used for series {textstyle sum _{n=0}^{infime }un_{n}} whose terms {style d'affichage a_{n}} are not numbers but rather elements of an arbitrary abelian topological group. Dans ce cas, instead of using the absolute value, the definition requires the group to have a norm, which is a positive real-valued function {style de texte |cdot |:Gto mathbb {R} _{+}} on an abelian group {style d'affichage G} (written additively, with identity element 0) tel que: The norm of the identity element of {style d'affichage G} est zéro: {style d'affichage |0|=0.} Pour chaque {displaystyle xin G,} {style d'affichage |X|=0} implique {displaystyle x=0.} Pour chaque {displaystyle xin G,} {style d'affichage |-X|=|X|.} Pour chaque {style d'affichage x,yin G,} {style d'affichage |x+y|leq |X|+|y|.} Dans ce cas, the function {displaystyle d(X,y)=|x-y|} induces the structure of a metric space (a type of topology) sur {displaystyle G.} Alors, un {style d'affichage G} -valued series is absolutely convergent if {textstyle sum _{n=0}^{infime }|un_{n}|0,} there exists {displaystyle N} such {textstyle left|sum _{i=m}^{n}left|a_{i}right|right| =sum _{i=m}^{n}|a_{i}|mgeq N.} But {textstyle {big |}sum _{i=m}^{n}a_{i}{big |}leq _{i=m}^{n}|a_{i}|,} so {textstyle left|sum _{i=m}^{n}a_{i}right|mgeq N,} exactly {textstyle sum a_{i}.} above result easily generalized {displaystyle (X,|,cdot ,|).} {textstyle sum x_{n}} {displaystyle X.} As {textstyle sum _{k=1}^{n}|x_{k}|} numbers, {displaystyle varepsilon >0} large enough natural {displaystyle m>n} holds: {displaystyle left|sum _{k=1}^{m}|x_{k}|-sum _{k=1}^{n}|x_{k}|right| =sum _{k=n+1}^{m}|x_{k}|kappa _{varepsilon }&quad sum _{n=N}^{infime }|un_{n}|<{tfrac {varepsilon }{2}}\{text{ for all }}N>lambda _{varepsilon }&quad left|somme _{n=1}^{N}un_{n}-Aright|<{tfrac {varepsilon }{2}}end{aligned}}} Let {displaystyle {begin{aligned}N_{varepsilon }&=max left{kappa _{varepsilon },lambda _{varepsilon }right}\M_{sigma ,varepsilon }&=max left{sigma ^{-1}left(left{1,ldots ,N_{varepsilon }right}right)right}end{aligned}}} where {displaystyle sigma ^{-1}left(left{1,ldots ,N_{varepsilon }right}right)=left{sigma ^{-1}(1),ldots ,sigma ^{-1}left(N_{varepsilon }right)right}} so that {displaystyle M_{sigma ,varepsilon }} is the smallest natural number such that the list {displaystyle a_{sigma (0)},ldots ,a_{sigma left(M_{sigma ,varepsilon }right)}} includes all of the terms {displaystyle a_{0},ldots ,a_{N_{varepsilon }}} (and possibly others). Finally for any integer {displaystyle N>M_{sigma ,varepsilon }} laisser {style d'affichage {commencer{aligné}JE_{sigma ,varepsilon }&=left{1,ldots ,Nright}setminus sigma ^{-1}la gauche(la gauche{1,ldots ,N_{varepsilon }droit}droit)\S_{sigma ,varepsilon }&=min sigma left(JE_{sigma ,varepsilon }droit)=min left{sigma (k) : kin I_{sigma ,varepsilon }droit}\L_{sigma ,varepsilon }&=max sigma left(JE_{sigma ,varepsilon }droit)=max left{sigma (k) : kin I_{sigma ,varepsilon }droit}\fin{aligné}}} pour que {style d'affichage {commencer{aligné}la gauche|somme _{iin I_{sigma ,varepsilon }}un_{sigma (je)}droit|&leq sum _{iin I_{sigma ,varepsilon }}la gauche|un_{sigma (je)}droit|\&leq sum _{j=S_{sigma ,varepsilon }}^{L_{sigma ,varepsilon }}la gauche|un_{j}droit|&&{texte{ puisque }}JE_{sigma ,varepsilon }subseteq left{S_{sigma ,varepsilon },S_{sigma ,varepsilon }+1,ldots ,L_{sigma ,varepsilon }droit}\&leq sum _{j=N_{varepsilon }+1}^{infime }la gauche|un_{j}droit|&&{texte{ puisque }}S_{sigma ,varepsilon }geq N_{varepsilon }+1\&<{frac {varepsilon }{2}}end{aligned}}} and thus {displaystyle {begin{aligned}left|sum _{i=1}^{N}a_{sigma (i)}-Aright|&=left|sum _{iin sigma ^{-1}left({1,dots ,N_{varepsilon }}right)}a_{sigma (i)}-A+sum _{iin I_{sigma ,varepsilon }}a_{sigma (i)}right|\&leq left|sum _{j=1}^{N_{varepsilon }}a_{j}-Aright|+left|sum _{iin I_{sigma ,varepsilon }}a_{sigma (i)}right|\&0,{texte{ il existe }}M_{sigma ,varepsilon },{texte{ pour tous }}N>M_{sigma ,varepsilon }quad left|somme _{je=1}^{N}un_{sigma (je)}-Aright|

Si vous voulez connaître d'autres articles similaires à Absolute convergence vous pouvez visiter la catégorie Integral calculus.

Monter

Nous utilisons nos propres cookies et ceux de tiers pour améliorer l'expérience utilisateur Plus d'informations