The rules that let you find limits of sequences without going back to epsilon each time, from limit laws and the squeeze theorem to subsequences and the Stolz-Cesàro theorem.
Checking every sequence limit straight from the definition would be slow and tiring. In practice you break a sequence into pieces whose limits you already know, trap it between two simpler sequences, or relate it to a function you can handle with ordinary calculus. The theorems below are what make those moves legal.
Key terms
Uniqueness of limits
Definition A sequence (an) converges to L if for every ε>0 there is an integer N such that ∣an−L∣<ε for all n≥N. If (an) converges to L and also to M, then L=M.
In plain words A sequence cannot settle down near two different numbers at once. So it makes sense to speak of the limit and write limn→∞an=L.
Example Suppose an→L and an→M with L=M, and take half their distance as the tolerance,
ε=2∣L−M∣>0.
Far enough out, an is within ε of both, so
∣L−M∣≤∣L−an∣+∣an−M∣<2ε=∣L−M∣,
which is impossible. This is also why an=(−1)n has no limit. Its terms keep landing on both 1 and −1, which are 2 apart, so they cannot all stay within 21 of a single number.
Limit laws
Definition Suppose an→L and bn→M, and let c be a constant. Then
an±bn→L±M,can→cL,anbn→LM,
and if M=0 and bn=0 for every n, then
bnan→ML.
Also anp→Lp for every positive integer p.
In plain words Limits of sequences respect arithmetic, as long as every piece has a finite limit and you never divide by a limit of zero.
Example Divide the top and bottom by the highest power of n.
2n2−n3n2+1=2−n13+n21→2−03+0=23.
The denominator 2−n1 is never zero for n≥1 and has limit 2=0, so the quotient law applies.
Comparison of limits
Definition If an→L, bn→M and an≤bn for all n≥N for some N, then L≤M.
In plain words An inequality between terms survives in the limit, but only as ≤. Strict inequalities between terms can turn into equality.
Example For every n≥1 we have 0<n1, yet limn1=0. So the strict inequality 0<n1 only gives 0≤limn1, and here equality happens.
Squeeze theorem
Definition If an≤bn≤cn for all n≥N and
n→∞liman=n→∞limcn=L,
then limn→∞bn=L. A useful special case is that ∣bn∣→0 implies bn→0.
In plain words Trap the hard sequence between two easy ones that head to the same place. It has no room to go anywhere else.
Example Since −1≤sinn≤1, for n≥1
−n1≤nsinn≤n1.
Both outer sequences go to 0, so nsinn→0. Notice that sinn itself has no limit. The squeeze works anyway.
Continuous functions of sequences
Definition If an→L, every an lies in the domain of f, and f is continuous at L, then
n→∞limf(an)=f(n→∞liman)=f(L).
In plain words A continuous function lets you pass the limit inside. Find where the inside goes, then apply f.
Example The inside sequence n4n+1=4+n1 tends to 4, and x is continuous at 4, so
n4n+1→4=2.
In the same way e1/n→e0=1.
Subsequences
Definition A subsequence of (an) is a sequence (ank)k≥1 where n1<n2<n3<⋯ are positive integers. If an→L, then every subsequence also converges to L. Moreover an→L if and only if both a2k→L and a2k−1→L.
In plain words Picking out some of the terms, in order, cannot change where a convergent sequence is headed. So if two subsequences head to different places, the whole sequence diverges.
Example Let an=(−1)nn+1n. The even and the odd terms are
a2k=2k+12k→1,a2k−1=−2k2k−1→−1.
Two subsequences with different limits mean (an) diverges.
Sequential criterion for function limits
Definition Let f be defined on an open interval containing a, except possibly at a. Then limx→af(x)=L if and only if f(xn)→L for every sequence (xn) in that interval with xn=a and xn→a. For limits at infinity there is a one way version. If limx→∞f(x)=L and an=f(n) for every integer n≥1, then an→L.
In plain words Function limits and sequence limits talk to each other. You can borrow tools like l’Hôpital’s rule by passing to a function of a real variable, and you can disprove a function limit by finding two sequences that disagree.
Example To find limnlnn, look at f(x)=xlnx. By l’Hôpital’s rule
x→∞limxlnx=x→∞lim11/x=0,
so nlnn→0. In the other direction, the limit of sinx1 as x→0 does not exist. Both sequences
xn=nπ1andyn=2nπ+π/21
tend to 0 and never equal 0, but sinxn1=0 for every n while sinyn1=1 for every n.
The converse of the one way version fails. The sequence sin(nπ)=0 converges, but sin(πx) has no limit as x→∞.
Stolz-Cesàro theorem
Definition Let (an) and (bn) be sequences of real numbers where (bn) is strictly increasing and bn→∞. If
n→∞limbn+1−bnan+1−an=L,
where L is a real number, ∞ or −∞, then
n→∞limbnan=L.
In plain words It is a discrete version of l’Hôpital’s rule. To find the limit of a quotient, look at how much the top and bottom change from one step to the next.
Example Take an=12+22+⋯+n2 and bn=n3, which is strictly increasing to ∞. Then
bn+1−bnan+1−an=3n2+3n+1(n+1)2→31,
and therefore
n312+22+⋯+n2→31.
The values 0.385, 0.33835 and 0.3338335 at n=10, 100 and 1000 agree.
Check yourself
Find limn→∞(n2+n−n).
Show the answer
Multiply and divide by the conjugate n2+n+n to get
n2+n−n=n2+n+nn=1+1/n+11.
By the limit laws and continuity of the square root this tends to 21.
Show that ncos(n2)→0.
Show the answer
Since −1≤cos(n2)≤1, for n≥1 we have
−n1≤ncos(n2)≤n1,
and both bounds go to 0. The squeeze theorem gives the limit 0.
Use the Stolz-Cesàro theorem to find limn21+2+⋯+n.
Show the answer
With bn=n2 strictly increasing to ∞, the quotient of differences is
May we count page visits to understand which courses and learning features are useful? Cloudflare collects visit and performance statistics, without your account details or answers. Your choice does not affect access. Details