0 Gy Sbn Seq forall n> N If lim ay Jim cp, then also lim by, ‘The theorem does not require that ay N Figure 1.6: For n > NV, the terms by is squeezed between ay, and by, ‘The next theor an immediate consequence of the Squeeze Theoren Theorem 5. Let {a,} be a sequence. If lim Jay] =0, then lim a =0. Proof. Observe that —an| 0. Hence, for all n > 1 OC tantn 1 tan! n Since lim 0=0= lim 4, by Squeeze Theorem, lim not neon nto Te Sec Use Squeeze Theorem to determine whether the sequence converge or diverge. If it converges, find the limit. cosn an = 1.3. Monotonic and Bounded Sequences We now introduce the concept of monotonicity and boundedness. ‘This will play a significant role in proving the convergence of real numbers. Definition 3: Monotone Sequence. A sequence is said to be (i) increasing if ay ay. for all n A sequence that is ther increasing or decreasing is called mono- tonic sequence. If the inequality were strict, then the sequence is said to be strictly monotonic. (Teme Lana cco Gab auc) Solution. From the first few terms of the sequence ‘The last inequality is true since the fraction on the left has larger denominator, hence, dy > Guys for all n > 1. Therefore, the sequence is indeed decreasing, . Here are the several tests from pre-calculus and introductory calculus that we can use to determine monotonicity of a sequence 1. Using the difference of successive terms. Given a sequence {an}16 Module 1. Infi ay Is Inereasiny

dn — Auer 2 0. ‘Using the ratio of successive terms, iven a sequence {ay} a aay is increasing => 1, Oat 3. Using the derivative of y = f(z). If a, = f(n) such that the fametion f is differentiable, then ay is increasing f(x) > 0. as, is decetoaning => f’(x) warp For @ > 1, the numerator is nevative while the dencminater is always negative hence. f"(@) . Therefore, the sequence is decreasing for all n > 5. . ‘Sometimes a sequence will not behave as monotonic at its first few terms (it oscillates up and down), but if distegarding a finitely many terms from the beginning of the sequence produces a sequence that is monotonic, then the original sequence is said to be eventually monotonic. What is important is that towards the end of the sequence, the terms eventually become either increasing or decreasing, Self-Check Test 1.7 Determine whether the sequence is increasing, decreasing or neither. If it is not mono- tonic, check if deleting the first few terms will result to a monotonic sequence.18 We now introduce the concept of boundedness Definition 4. A sequence {ay} is hounded above if ther ammmber M such that a, 1 The mumber m is called a lower bound of the sequence. 4 sequenee is said to be bounded if it is both hounded above and bounded below, (eee ecu oa 1. The sequence defined by a, = nis bounded belew because ay, > 0 for all n, but it is not bounded above, . The sequence defined by aq 78 hounded becatse 0 5, all the remaining tenis awe Kee as, Therefore, a5 = 625/120 is an upper bound, Mercentny since all the tems are positive. a, > 0 for alll n, Hence, the sequence is bounded, One important property of real number regarding boundedness is that it is complete inva seuse that it has ue gap or uo hole iu the umuber line, Given any two teal uamubers rand s, we can always find another real number # in between, that iss 7 625/120, Hence 625/120 is the smallest of all the upper bounds. and is therefore. the least upper hound. On the other hand. no lower bound of the sequence can be positive since the terms of th sequence are positive mmbers approaching, as m —¥ 00, Hence all the lower bounds a smaller than 0, In other words. 0 is the Inngest of all the lower bounds, and is therefon the greatest lower bound. We now connect boundedness and monotonicity. Consider an eventually increasin sccquenee {ay}. that is, there exists an index N’ such that ay Sayer forall n> Then there are two possibilities, either L. the sequence will continue to increase without bound. in which case lim ay = 00: or 2, the sequence has an upper bound, sav M. in which case the infinitely many terms that are eventually increasing (hut cannot go further above M) are forced to crow together and converge to a certain limit Z with L M, This can be summarized in the next theorem, Theorem 6: Bounded Monotonic Sequence, \ bounded monotenic sequence is convergent, Proof, We will prove the case when the sequence is increasing, The case when a sceence is decreasing can be a wood exercike of your proving skill, Suppose {an} is a bounded inercasing sequence, Let $ be the set containing all the terms of the sequence. that is, = {ay 2 n> 1} Since {aq} is hounded abexe. $ has an upper bound. By the Cempleteness Axiom, $1 has a least upper bound. sav L, We claim that this L is the limit of the sequence which makes the sequence convergent. Let € > O be wiven, Since Lis the least upperbound. the number L ~e (a number Which is smaller than Z) cannct be an upperbound for $. This means that Shas an Joment bigerr than L—e, that is, ay > L—e for some subscript N (see Figure 1 So what can you sav about the succeeding terms after ay’ Since {ay} is increasing. ay L-' Loa -e |L—ag| a a1.3 Monotonic and Bounded Sequences 21 1, Show that a, > V3 for n > 2, 2, Show that {a,} is eventually decreasing, 3. Show that {ay} is convergent and find its limi. Solution, So that we have a general idea of what's happening on the terms of tly sequence, we ise a calculater to find the fitst few terms of the sequence, Sensis) CS 1 m=1 100000000 2 m=} ‘ocen0004 3 aa} eH0000¢ I a4} 32112857 5 as} 2205081C 6 25=3 [4+ echo 725 = 5 | ogssrre + rosrssorrjaussssrre| = }saserrer et? “The values of zp and 7 are actualy not equal, but the calculator used can no longer show digits ‘beyond the 10th decimal place. Using the table, we can already hypothesize that (1). aq > v3 = 1.782050807569 for n> 2, (2) the sequence is decreasing and (2) ay 9 V3 as n> 00. We now proxe thes eins. 1, Consider the function 1o)-3(2+8)- We shoul be able to recoenize that this is a rational function whose graph has vertical asvimptote 2x = 0 and oblique asvinptote y = dx (sce Fieure 18). For > 0. the staph has a mininmm value at = VB which is f(v3) way? 2v3 V3. Figure 1.8: y = (x? +.3)/2r for x > 0. Since the terms @qj1 = § (ay + 2) of the sequence are points en the eraph of f(x) =4 (2 +4), it follows that a, > V3 for all n > 2. 2. We compute the ratio ans1/an to determine the monotonicity of the sequences W haw From the previous result, ay > V3 for all n > 2, it follows that 35a a+ a-w +a22 Module 1. Infinite Sequences Hence an+1/a, 2, By (1), all the terms of the sequence a, > V3 for all n > 2 and by (2), an is decreasing starting from n = 2. Then the sequence is bounded above by ay = 2 and below by a; = 1. By the Bounded Monotone Sequence Theorem, {an} is convergent. So how do we find the limit of the sequence? Knowing that the sequence converge to a limit, say to a posi can assert that as n > 00, dy > Land dyy1 > L. It follows that Jim aye = ] + ae = b= = Weal? +3 = P=3 = L=V3 (since L > 0) . Let {a,} be the sequence define recursively by . Show that a a= V2, and anya = V2 aq for n> 1 List the first three terms of the sequence. . Show that ay 1 2 By — a2 = (2 ay)(1+ an) for n> 1 Use the previous results to show that {aq} is strictly increasing. Hint: Use the fact that for positive real numbers x and y, if x? — y? > 0 then x —y > 0. Show that {a,} converge and find its limit. Solution on page 79