TheoremBase

Induction gives the bounds and the monotonicity through the equivalence between an inequality of nonnegative numbers and the inequality of their squares; the monotone convergence theorem gives a limit LL, and passing to the limit in an+12=2+ana_{n+1}^2=2+a_n forces L=2L=2.

Proof

Throughout we use that, for every n∈Nn\in\mathbb{N}, the number an+1a_{n+1} satisfies 0≀an+10\le a_{n+1} and an+12=2+ana_{n+1}^2=2+a_n, which is what the definition of 2+an\sqrt{2+a_n} says.

Step 1: 0≀an<20\le a_n<2 for every n∈Nn\in\mathbb{N}. We argue by the principle of induction. For n=1n=1 we have a1=1a_1=1, so 0≀a10\le a_1 and a1<2a_1<2. Suppose 0≀an<20\le a_n<2. Then 0≀an+10\le a_{n+1}, and

an+12=2+an<2+2=4=2β‹…2.a_{n+1}^2=2+a_n<2+2=4=2\cdot 2 .

Since 0≀an+10\le a_{n+1} and 0≀20\le 2, clause 1 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field converts an+12<22a_{n+1}^2<2^2 into an+1<2a_{n+1}<2. Hence 0≀an+1<20\le a_{n+1}<2, completing the induction.

Step 2: an≀an+1a_n\le a_{n+1} for every n∈Nn\in\mathbb{N}. Fix nn. By Step 1, 0≀an<20\le a_n<2, so anβˆ’2<0a_n-2<0 and 0<1≀an+10<1\le a_n+1. By clause 4 of Elementary Order Arithmetic in an Ordered Field, anβˆ’2<0a_n-2<0 gives 0<βˆ’(anβˆ’2)0<-(a_n-2); by clause 5 of that lemma, 0<(βˆ’(anβˆ’2))(an+1)0<\bigl(-(a_n-2)\bigr)(a_n+1); and since (βˆ’(anβˆ’2))(an+1)=βˆ’((anβˆ’2)(an+1))\bigl(-(a_n-2)\bigr)(a_n+1)=-\bigl((a_n-2)(a_n+1)\bigr) by Elementary Arithmetic in an Ordered Field, clause 4 applied again yields

(anβˆ’2)(an+1)<0.(a_n-2)(a_n+1)<0 .

Expanding the left side using Elementary Arithmetic in an Ordered Field gives an2βˆ’anβˆ’2<0a_n^2-a_n-2<0, that is,

an2<2+an=an+12.a_n^2<2+a_n=a_{n+1}^2 .

Since 0≀an0\le a_n and 0≀an+10\le a_{n+1}, clause 1 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field gives an<an+1a_n<a_{n+1}; in particular an≀an+1a_n\le a_{n+1}.

Step 3: convergence. Let A={an:n∈N}A=\{a_n:n\in\mathbb{N}\}. By Step 1 every element of AA is at most 22, so AA is nonempty and bounded above, and by Step 2 the sequence satisfies an≀an+1a_n\le a_{n+1} for every nn. By clause 1 of A Bounded Monotone Sequence of Real Numbers Converges, L=sup⁑AL=\sup A exists and (an)(a_n) converges to LL. Since 22 is an upper bound for AA and LL is the least upper bound, L≀2L\le 2; and since a1=1a_1=1 lies in AA and LL is an upper bound, 1≀L1\le L.

Step 4: the shifted sequence converges to LL as well. Put bn=an+1b_n=a_{n+1} for n∈Nn\in\mathbb{N}. By the index-shift property recorded in The Real Line: Standing Notation and Background for Calculus on Intervals (vii), (bn)(b_n) converges to LL.

Step 5: identifying LL. By clause 2 of Arithmetic of Limits of Real Sequences applied to the sequence (bn)(b_n) and itself, the sequence (bn bn)=(bn2)(b_n\,b_n)=(b_n^2) converges to Lβ‹…L=L2L\cdot L=L^2. On the other hand, the constant sequence with value 22 converges to 22 by The Real Line: Standing Notation and Background for Calculus on Intervals (vii), so by clause 1 of Arithmetic of Limits of Real Sequences the sequence (2+an)(2+a_n) converges to 2+L2+L. But bn2=an+12=2+anb_n^2=a_{n+1}^2=2+a_n for every nn, so the two sequences are equal, and by the uniqueness of limits in clause 1 of Uniqueness of Limits and Boundedness of Convergent Real Sequences,

L2=2+L.L^2=2+L .

Rearranging with Elementary Arithmetic in an Ordered Field gives L2βˆ’Lβˆ’2=0L^2-L-2=0, that is, (Lβˆ’2)(L+1)=0(L-2)(L+1)=0. A field has no zero divisors, so Lβˆ’2=0L-2=0 or L+1=0L+1=0, that is, L=2L=2 or L=βˆ’1L=-1. By Step 3, 1≀L1\le L, so Lβ‰ βˆ’1L\ne -1. Therefore L=2L=2, that is,

lim⁑nβ†’βˆžan=2.\lim_{n\to\infty}a_n=2 .

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…