Math for Non-Geeks/Divergence to infinity/Rules

From testwiki
Jump to navigation Jump to search

{{#invoke:Math for Non-Geeks/Seite|oben}} In the last article, we mentioned that some rules for limit calculation carry through to improper convergent sequences, and some don't. For instance, if a sequence (an) (improperly) converges to and a second sequence (bn) (properly) converges to 0, one cannot make any statement about the convergence of their product! Math for Non-Geeks: Template:Aufgabe

Rules for computing limits of improperly converging sequences

Which calculation rules for limits of convergent sequences can be carried over to improper convergence? The answer is: almost all of them, but only if certain conditions hold!

Product rule

Suppose that (an) is a sequence with limnan= . What will happen to the product (anbn) ? The case limnbn=0 definitely causes trouble, meaning that we cannot make any statement about convergence or divergence of the product.

Case 1: limnbn=. Intuitively, = so we expect limnanbn= . This assertion only needs to be mathematically proven: Template:Math Let S be given. Since limnan= we can find an N1 with an|S| for all nN1. Analogously, since limnbn= there is an N2 with bn|S| for all nN2. Whenever nN=max{N1,N2} we therefore have Template:Math So, indeed limnanbn=.

Case 2: limnbn=. Intuitively, = so we expect limnanbn= . What we need to show for a mathematical proof is: Template:Math So let again S be given. Since limnan= there is an N1 with an|S| for all nN1. Analogously, since limnbn= there is an N2 with bn|S| for all nN2. Now, for all nN=max{N1,N2} we have Template:Math And indeed there is limnanbn=.

Case 3: limnbn=b>0. Intuitively, = so we again make a guess limnanbn=. The proof could be done as the two examples above. However, this time we will vary it a bit, to make it not too boring: Template:Math Let S be given. Since limnbn=b , for each ϵ>0 there is an N1 with |bnb|<ϵ for all nN1. We set ϵ=b2. Then there is nN1: |bnb|<b2, which especially includes bn>b2>0. Since limnan= there is some N2 with an|S|b2=2|S|b0 for all nN1. Now for nN=max{N1,N2} we have Template:Math And hence limnanbn=.

Case 4: limnbn=b<0. Here, limnanbn=. Math for Non-Geeks: Template:Aufgabe

Those four cases can also be concluded into one statement. We introduce a practical extension of the real numbers: To the set , we add the elements ± which leads to the bigger set ={±}. Math for Non-Geeks: Template:Satz

Sum rule

Let again (an) be a sequence with limnan=. What can we say about the limit of a sum (an+bn) ? For finite bn, the limit will stay unchanged, as intuitively +c=. Similarly +=. The critical case is limnbn= , as is not well-defined. In fact, this case does not allow for any statement about convergence or divergence of the sum (an+bn). As an example,

  • For an=n and bn=n there is limnan+bn=limn0=0.
  • For an=2n and bn=n there is limnan+bn=limnn=.

We therefore exclude the case limnbn= and consider all other cases:

Case 1: limnbn=. We expect limnan+bn= Mathematically, we need to prove: Template:Math Let S be given. Since limnan= there is an N1 with anS2 for all nN1. Analogously, since limnbn= there is an N2 with bnS2 for all nN2. Hence, for all nN=max{N1,N2} we have

Template:Math

And indeed limnan+bn=.

Case 2: limnbn=b. We also expect limnan+bn=. Mathematically, we need to prove: Template:Math Let S be given. Since limnbn=b for each ϵ>0 we can find an N2 with |bnb|<ϵ for all nN2. This includes the case ϵ=1. Hence, bn>b1 for all nN2. Since limnan= there is also an N1 with anSb+1 for all nN1. Hence, for any nN=max{N1,N2} we have

Template:Math

And we get the desired result limnan+bn=.

Both cases can be concluded in a theorem:

Math for Non-Geeks: Template:Satz

Inversion

This rule is also quite intuitive: Let (an) be a sequence with an0 for all n and limnan= or limnan=, then (1an) formally converges to 1=0 and should hence be a null sequence. Is this really mathematically true?

Case 1: limnan=. We need to show Template:Math Let ϵ>0 be given. Since limnan= for any S=1ϵ+1 there is an N, such that nN there is an1ϵ+1. Hence, nN there is Template:Math Solimnan=0.

Case 2: limnan=. Math for Non-Geeks: Template:Aufgabe

We conclude these findings in a theorem: Math for Non-Geeks: Template:Satz

Math for Non-Geeks: Template:Frage

The question is now: can we define a "converse of the inversion rule" which holds under more special assumptions? The sequence ((1)nn) is not diverging to or because it keeps changing presign, so there is a subsequence of it converging to and a subsequence converging to . We can avoid this by forbidding a change of presign in (an). It should also not be too bad if the change of presign is allowed again on finitely many elements, since a manipulation on finitely many elements never changes convergence properties.

Case 1: Let (an) be a sequence with limn1an=0, all sequence elements being 0 and all but finitely many sequence elements being positive. Then, intuitively limnan=1+0=. For a mathematical proof, we need to show that Template:Math Let S be given. Since (1an) is a null sequence, for ϵ=1|S|>0 we can find an N~ with |1an|<1|S| for all nN~. Since almost all elements of (an) are positive, there is an NN~ with |1an|=1an<1|S| for all nN. Therefore an|S|S for all nN. So we get limnan=.

Case 2: Let now (an) v limn1an=0, all sequence elements being 0 but this time, almost all of them are negative. Math for Non-Geeks: Template:Aufgabe

The converse of the inversion rule is also concluded into a theorem. Math for Non-Geeks: Template:Satz

Math for Non-Geeks: Template:Beispiel

Quotient rule

The inversion rule is an example of a quotient anbn=1bn of sequences (an) and (bn) wit constant an=1. Now, we generalize to quotients (anbn) of any sequences (an) and (bn) with bn0 for all n .

First, we consider limnbn= . At this point, we exclude the cases limnan=± , since ± is ill-defined. Let limnan=a. Then, formally limnanbn=a=0. To verify this mathematically, we need to show Template:Math Let ϵ>0 be given. Since there is convergence limnan=a , the sequence (an) must be bounded, i.e. there is a K>0 with |an|K for all n. Now since limnbn=, the inversion rule implies limn1bn=0. So there is an N with |1bn|<ϵK for all nN. Hence, nN there is: Template:Math And we have convergence limnanbn=0.

The case limnbn= and limnan=a also leads to limnanbn=0 by the same argument.

We conclude Math for Non-Geeks: Template:Satz

Next, we let the enumerator diverge as limnan=. The case limnbn=± again leads to the ill-defined expression ± and will not be not considered at this point.

Case 1: limnbn=b>0. Here, we assert limnanbn=b=. : Template:Math Let S be given. Since (bn) converges to b>0, there must be an N1, such that bn32b for all nN1. Since limnan= there is also an N2 with an32b|S| for all nN2. Hence, for all nN=max{N1,N2}, there is: Template:Math So we have convergence limnanbn=.

Case 2: limnbn=0 with almost all bn being positive. Here, we assert limnanbn=+0=. A mathematical proof requires showing Template:Math Let S be given. Since (bn) converges to 0 and almost all elements are positive, there must be an N1 with bn12 for all nN1. Since limnan= there is also an N2 with an12|S| for all nN2. So for all nN=max{N1,N2}, there is: Template:Math And again, we have convergence limnanbn=.

Math for Non-Geeks: Template:Aufgabe

All 4 cases are concluded in a theorem Math for Non-Geeks: Template:Satz

Direct comparison

Intuitively, if (xn) is given and some "smaller" sequence (yn) diverges to , then also the "bigger" (xn) must tend to . This should still hold true if "(xn) is bigger than (yn)" almost everywhere. Mathematically, we need to show Template:Math So let S be given. Since limnyn= there is an N~ with ynS for all nN~. Since xnyn for all but finitely many n there is an NN~ with xnynS for all nN. So indeed, limnxn=.

We conclude this in a theorem: Math for Non-Geeks: Template:Satz

Math for Non-Geeks: Template:Beispiel

Of course, a similar statement holds true for xnyn and limnyn=. Then also limnxn= . This can easily seen by considering the sequences (|xn|)n and (|yn|)n.

{{#invoke:Math for Non-Geeks/Seite|unten}}

Template:BookCat