← 2026 Paper 2

UPSC 2026 Maths Optional Paper 2 Q2b — Step-by-Step Solution

15 marks · Section A

Sequences · Real Analysis · asked 3× in 14 yrs · Read the full method →

Question

Prove that the sequence (un)(u_n) defined by the recursion formula un+1=7+unu_{n+1} = \sqrt{7 + u_n}, u1=7u_1 = \sqrt{7}, converges to the positive root of x2−x−7=0x^2 - x - 7 = 0.

Technique

The trigger is a recursion un+1=f(un)u_{n+1}=f(u_n) with ff increasing and a fixed point: the route is the monotone convergence theorem — bounded above ++ increasing ⇒\Rightarrow convergent — and the correct bound to induct on is the fixed point itself, i.e. the positive root LL of x2−x−7=0x^{2}-x-7=0, because L=7+LL=\sqrt{7+L} makes the induction step an identity rather than an estimate. Convergence must be established first; only then may one pass to the limit in the recursion. Squaring the recursion to un+12=7+unu_{n+1}^{2}=7+u_n before taking limits removes even the need to invoke continuity of ⋅ \sqrt{\cdot\,}.

Solution

Step 0 — The target number, named once and for all.

The roots of x2−x−7=0x^{2}-x-7=0 are 1±292\dfrac{1\pm\sqrt{29}}{2}. Their product is −7<0-7<0, so they have opposite signs; there is therefore exactly one positive root, namely

L=1+292  (>0),the other root being M=1−292<0  (as 29>1).L=\frac{1+\sqrt{29}}{2}\;(>0),\qquad\text{the other root being } M=\frac{1-\sqrt{29}}{2}<0 \ \ (\text{as }\sqrt{29}>1).

Two properties of LL will be used repeatedly:

L2=L+7,equivalentlyL=7+L(positive square root, legitimate since L>0).(0.1)L^{2}=L+7,\qquad\text{equivalently}\qquad L=\sqrt{7+L}\quad(\text{positive square root, legitimate since }L>0). \tag{0.1}

Numerically L=1+292≈3.1926L=\tfrac{1+\sqrt{29}}{2}\approx 3.1926, while u1=7≈2.6458u_1=\sqrt7\approx2.6458.

Step 1 — The sequence is well defined and positive.

u1=7>0u_1=\sqrt7>0. If un>0u_n>0 then 7+un>7>07+u_n>7>0, so 7+un\sqrt{7+u_n} exists as a real number and un+1=7+un>7>0u_{n+1}=\sqrt{7+u_n}>\sqrt7>0. By induction un>0u_n>0 for all n≥1n\ge1, and the recursion never leaves the reals. (Throughout, ⋅ \sqrt{\cdot\,} denotes the non-negative square root.)

Step 2 — (un)(u_n) is bounded above by LL.

Claim: un<Lu_n<L for all n≥1n\ge1.

Base n=1n=1: since L>0L>0,

u1=7=7+0<7+L=(0.1)L.u_1=\sqrt7=\sqrt{7+0}<\sqrt{7+L}\overset{(0.1)}{=}L .

Induction step: suppose un<Lu_n<L. The function t↦7+tt\mapsto\sqrt{7+t} is strictly increasing on [0,∞)[0,\infty), so

un+1=7+un<7+L=(0.1)L.u_{n+1}=\sqrt{7+u_n}<\sqrt{7+L}\overset{(0.1)}{=}L .

By induction un<Lu_n<L for every nn. In particular (un)(u_n) is bounded above, and 0<un<L0<u_n<L for all nn. (2.1)\qquad(2.1)

(This is exactly why LL, and not some convenient round number, is the bound to choose: (0.1)(0.1) turns the induction step into an equality, so nothing has to be estimated.)

Step 3 — (un)(u_n) is monotonically increasing.

Base: u2=7+u1=7+7u_2=\sqrt{7+u_1}=\sqrt{7+\sqrt7} and u1=7=7+0u_1=\sqrt7=\sqrt{7+0}. Since 7>0\sqrt7>0 we have 7+7>7+07+\sqrt7>7+0, and ⋅ \sqrt{\cdot\,} is strictly increasing, so

u2>u1.u_2>u_1 .

Induction step: suppose un+1>unu_{n+1}>u_n. Then 7+un+1>7+un7+u_{n+1}>7+u_n, and applying the strictly increasing map t↦tt\mapsto\sqrt t,

un+2=7+un+1>7+un=un+1.u_{n+2}=\sqrt{7+u_{n+1}}>\sqrt{7+u_n}=u_{n+1}.

Hence un+1>unu_{n+1}>u_n for all n≥1n\ge1: the sequence is strictly increasing. (3.1)\qquad(3.1)

Cross-check by the quadratic (independent of the induction above). From the recursion,

un+12−un2=(7+un)−un2=−(un2−un−7)=− (un−L)(un−M).u_{n+1}^{2}-u_n^{2}=(7+u_n)-u_n^{2}=-\bigl(u_n^{2}-u_n-7\bigr)=-\,(u_n-L)(u_n-M).

By (2.1)(2.1), M<0<un<LM<0<u_n<L, so (un−L)<0(u_n-L)<0 and (un−M)>0(u_n-M)>0, making the right side positive. Thus un+12>un2u_{n+1}^{2}>u_n^{2}, and as both terms are positive, un+1>unu_{n+1}>u_n. ✓ (This version also displays the mechanism: the sign of q(x)=7+x−x2q(x)=7+x-x^{2} on the interval (M,L)(M,L) is what drives the sequence upward, and it is the vanishing of qq at LL that stops it.)

Step 4 — Apply the Monotone Convergence Theorem.

Theorem (Monotone Convergence). A monotonically increasing sequence of real numbers which is bounded above converges, and its limit is its supremum.

By (3.1)(3.1) the sequence is increasing and by (2.1)(2.1) it is bounded above by LL. Therefore (un)(u_n) converges. Write

ℓ=lim⁡n→∞un=sup⁡n≥1un.\ell=\lim_{n\to\infty}u_n=\sup_{n\ge1}u_n .

Since un<Lu_n<L for all nn, the supremum satisfies ℓ≤L\ell\le L; since the sequence increases from u1u_1, we also get

ℓ≥u1=7>0.(4.1)\ell\ge u_1=\sqrt7>0. \tag{4.1}

This is the load-bearing step, and it must come before Step 5. The classical error is to begin by writing “let ℓ=lim⁡un\ell=\lim u_n, then ℓ=7+ℓ\ell=\sqrt{7+\ell}”, which presumes precisely what is to be proved. The sequence un+1=1+un2u_{n+1}=1+u_n^{2}, u1=1u_1=1, satisfies the analogous fixed-point equation ℓ=1+ℓ2\ell=1+\ell^{2} with no real solution at all — because it diverges. Existence of the limit is not a formality; it is the theorem.

Step 5 — Pass to the limit in the recursion (now legitimate).

Square the recursion — this keeps everything polynomial and avoids appealing to the continuity of ⋅ \sqrt{\cdot\,}:

un+12=7+un(n≥1).u_{n+1}^{2}=7+u_n\qquad(n\ge1).

The sequence (un+1)n≥1(u_{n+1})_{n\ge1} is a tail of (un)(u_n) and hence converges to the same limit ℓ\ell. By the algebra of limits, un+12→ℓ2u_{n+1}^{2}\to\ell^{2} and 7+un→7+ℓ7+u_n\to 7+\ell. Limits of equal sequences are equal, so

ℓ2=7+ℓ,i.e.ℓ2−ℓ−7=0.\ell^{2}=7+\ell,\qquad\text{i.e.}\qquad \ell^{2}-\ell-7=0 .

Thus ℓ\ell is a root of x2−x−7=0x^{2}-x-7=0, so ℓ∈{L,M}\ell\in\{L,M\}.

Step 6 — Identify the root: it is the positive one.

By (4.1)(4.1), ℓ≥7>0\ell\ge\sqrt7>0, whereas M=1−292<0M=\dfrac{1-\sqrt{29}}{2}<0. Hence ℓ≠M\ell\neq M, and therefore

ℓ=L=1+292,\ell=L=\frac{1+\sqrt{29}}{2},

the positive root of x2−x−7=0x^{2}-x-7=0. (Two independent reasons exclude MM: every un>0u_n>0 so the limit cannot be negative; and ℓ≤L\ell\le L from Step 4 combined with ℓ∈{L,M}\ell\in\{L,M\} and ℓ>0\ell>0 pins ℓ=L\ell=L at once.)

Therefore (un)(u_n) converges, and

lim⁡n→∞un=1+292≈3.1926.■\lim_{n\to\infty}u_n=\frac{1+\sqrt{29}}{2}\approx 3.1926 .\qquad\blacksquare

Step 7 — Remark: the convergence is geometric (free extra credit).

Using A−B=A−BA+B\sqrt{A}-\sqrt{B}=\dfrac{A-B}{\sqrt A+\sqrt B} with A=7+LA=7+L, B=7+unB=7+u_n,

0<L−un+1=7+L−7+un=L−un7+L+7+un=L−unL+un+1.0<L-u_{n+1}=\sqrt{7+L}-\sqrt{7+u_n}=\frac{L-u_n}{\sqrt{7+L}+\sqrt{7+u_n}}=\frac{L-u_n}{L+u_{n+1}} .

Since un+1>u1=7u_{n+1}>u_1=\sqrt7 (Step 3) and L=7+L>7L=\sqrt{7+L}>\sqrt7, the denominator exceeds 27=28>52\sqrt7=\sqrt{28}>5, so

0<L−un+1<15 (L−un)⟹0<L−un<(15) n−1(L−u1)<(15) n−1,0<L-u_{n+1}<\tfrac15\,(L-u_n)\quad\Longrightarrow\quad 0<L-u_n<\Bigl(\tfrac15\Bigr)^{\,n-1}(L-u_1)<\Bigl(\tfrac15\Bigr)^{\,n-1},

the last step because L−u1=L−7<1L-u_1=L-\sqrt7<1: indeed q(x)=x2−x−7>0q(x)=x^{2}-x-7>0 for x>Lx>L, and q(1+7)=(8+27)−(1+7)−7=7>0q(1+\sqrt7)=(8+2\sqrt7)-(1+\sqrt7)-7=\sqrt7>0, so 1+7>L1+\sqrt7>L. So un→Lu_n\to L at least as fast as 5−n5^{-n} — which both re-proves convergence (by the sandwich theorem, independently of Step 4) and quantifies it.

Answer

  (un) is strictly increasing and bounded above by L, hence convergent, and lim⁡n→∞un=L=1+292, the positive root of x2−x−7=0.  \boxed{\;(u_n)\ \text{is strictly increasing and bounded above by }L,\ \text{hence convergent, and}\ \lim_{n\to\infty}u_n=L=\frac{1+\sqrt{29}}{2},\ \text{the positive root of } x^{2}-x-7=0.\;}
We post more of this — worked solutions, CSAT trap breakdowns, guide chapters — a few times a week on Telegram. Free, no sign-in. Join

This solution is part of the Maths Coverage Map — 14 years, mapped. Get the take-away PDF free.