MTH303 · Sample Midsem · 36 Marks
Sample Midsem — Complete Solutions
IISER Bhopal · Semester I 2026–27 · Time: 2 hours
- The paper carries 36 marks; the recorded score is out of 30, so marks beyond 30 count as bonus.
- You may use a theorem from class or the notes, provided you state it clearly.
- Throughout, \(\mathbb{N}=\{1,2,3,\dots\}\). Open, closed, and compact refer to subsets of \(\mathbb{R}\).
- Total 36 marks hain, lekin score 30 mein record hoga; 30 se upar ke marks bonus hain.
- Class ka theorem use kar sakte ho, bas theorem ka naam aur conditions clearly likho.
- Har proof mein main idea, justification aur exact conclusion clearly present karo.
Dotted keywords open their definition, Hinglish meaning, visualization, and properties.
Dotted keyword par click karke definition, Hinglish meaning, visual aur properties dekh sakte ho.
Questions 1–3 · 12 Marks
Sequences & Bounds
Define \(x_1=1\) and \(x_{n+1}=\sqrt{2+x_n}\) for \(n\ge1\). Prove that \((x_n)\) converges and find its limit.
Boundedness. We prove by induction that \(0<x_n<2\). It holds for \(x_1=1\). If \(0<x_n<2\), then \(0<x_{n+1}=\sqrt{2+x_n}<\sqrt4=2\). Hence the sequence is bounded above by \(2\).
Monotonicity. Since both terms are positive, \[x_{n+1}>x_n \iff 2+x_n>x_n^2 \iff (2-x_n)(1+x_n)>0.\] This holds because \(0<x_n<2\), so \((x_n)\) is increasing.
Limit. The Monotone Convergence Theorem gives \(x_n\to L\) with \(L\ge1\). Passing to the limit in \(x_{n+1}^2=2+x_n\) gives \(L^2=2+L\). Thus \(L=2\) or \(L=-1\); since \(L\ge1\), \(\boxed{L=2}\).
Sequence ko control karo. Induction se har term \(0\) aur \(2\) ke beech rehta hai. Agar \(x_n<2\), toh square root ke andar \(2+x_n<4\), isliye next term bhi \(2\) se chhota rahega.
Direction check karo. \(x_{n+1}>x_n\) ko square karke \((2-x_n)(1+x_n)>0\) milta hai. Dono factors positive hain, so sequence increasing bhi hai aur \(2\) se bounded bhi.
Limit identify karo. Monotone Convergence Theorem convergence guarantee karta hai. Recurrence mein limit dalne par \(L^2=L+2\) milta hai. Possible values \(2\) aur \(-1\) hain, lekin sequence positive hai; final limit \(\boxed{2}\) hai.
(a) Prove every Cauchy sequence is bounded. (b) If \(|x_{n+1}-x_n|\le2^{-n}\), prove \((x_n)\) converges.
(a) Use the Cauchy definition with \(\varepsilon=1\). There is \(N\) such that \(|x_n-x_m|<1\) for \(m,n\ge N\). Taking \(m=N\), we get \(|x_n|<|x_N|+1\) for \(n\ge N\). Therefore \[M=\max\{|x_1|,\dots,|x_{N-1}|,|x_N|+1\}\] bounds every term.
(b) If \(m>n\), then \[|x_m-x_n|\le\sum_{k=n}^{m-1}|x_{k+1}-x_k|\le\sum_{k=n}^{m-1}2^{-k}<2^{-n+1}.\] Given \(\varepsilon>0\), choose \(N\) with \(2^{-N+1}<\varepsilon\). Then \(m>n\ge N\) implies \(|x_m-x_n|<\varepsilon\). Thus \((x_n)\) is Cauchy and converges in \(\mathbb{R}\).
(a) Cauchy sequence ke late terms aapas mein close hote hain. \(\varepsilon=1\) lekar tail ko fixed term \(x_N\) ke around ek bounded box mein rakh do. Starting ke finitely many terms ka maximum alag le lo; dono bounds milkar poori sequence ko bound kar dete hain.
(b) \(x_n\) se \(x_m\) tak change ko consecutive jumps mein todte hain. Ye jumps geometric-series tail se bounded hain, aur tail \(0\) ki taraf jata hai. Isliye late terms arbitrarily close hain: sequence Cauchy hai. \(\mathbb R\) complete hai, so sequence converge karti hai.
For \(A=\{\frac1m-\frac1n:m,n\in\mathbb N\}\), find \(\sup A\) and \(\inf A\), prove your answer, and decide whether they are attained.
Supremum. For all \(m,n\), \(\frac1m-\frac1n<\frac1m\le1\), so \(1\) is an upper bound and is never attained. Given \(\varepsilon>0\), choose \(n\) with \(1/n<\varepsilon\). Taking \(m=1\) gives \(1-1/n\in A\) and \(1-1/n>1-\varepsilon\). Hence \(\sup A=1\).
Infimum. Swapping \(m\) and \(n\) negates an element, so \(A=-A\). Therefore \(\inf A=-\sup A=-1\). Neither bound is attained.
Upper edge. Expression \(1\) tak nahi pahunchta, kyunki positive \(1/n\) subtract ho raha hai. Lekin \(m=1\) fix karke \(n\) bada karo, toh \(1-1/n\), \(1\) ke arbitrarily paas aata hai. Isliye supremum \(1\) hai, maximum nahi.
Lower edge. \(m,n\) swap karne se sign reverse hota hai, so set zero ke around symmetric hai. Isliye infimum \(-1\) hai. Dono bounds sirf approach hote hain, attain nahi.
Questions 4–6 · 12 Marks
Topology & Compactness
(a) Prove an arbitrary intersection of closed sets is closed. (b) Is \(\mathbb Q\cap[0,1]\) compact?
(a) Let \(F=\bigcap_\alpha F_\alpha\), with every \(F_\alpha\) closed. De Morgan's law gives \(F^c=\bigcup_\alpha F_\alpha^c\). Each complement is open, and an arbitrary union of open sets is open. Hence \(F^c\) is open and \(F\) is closed.
(b) It is not compact. Take \(x=1/\sqrt2\in[0,1]\setminus\mathbb Q\). Every neighbourhood of \(x\) contains rationals by density of \(\mathbb Q\), so \(x\) is a limit point of \(\mathbb Q\cap[0,1]\) not contained in the set. The set is not closed and, by Heine–Borel, not compact.
(a) Closed sets ke complements open hote hain. Intersection ka complement union ban jata hai; arbitrary union of open sets open hota hai. Isliye original intersection closed hai.
(b) Interval ke andar irrational \(1/\sqrt2\) lo. Rationals dense hain, so rational points iske arbitrarily paas milte hain, par point khud set mein nahi. Set apna limit point miss karta hai, so closed nahi; Heine–Borel se compact bhi nahi.
Let \(K\subseteq\mathbb R\) be compact and \(F\subseteq\mathbb R\) closed. Without Heine–Borel, prove \(K\cap F\) is compact.
Use sequential compactness. Let \((x_n)\) be any sequence in \(K\cap F\). Compactness of \(K\) gives a subsequence \(x_{n_k}\to x\) with \(x\in K\). Every subsequence term lies in the closed set \(F\), so its limit lies in \(F\). Thus \(x\in K\cap F\). Every sequence in the intersection has a subsequence converging to a point of the intersection, so \(K\cap F\) is sequentially compact and hence compact.
Heine–Borel allowed nahi hai, so compactness ka sequence version use karo. Intersection ki koi sequence lo. K compact hone se convergent subsequence milti hai jiska limit K mein hai. Saare terms F mein bhi hain, aur F closed hai, so limit F ke andar rahega. Limit intersection mein mila; hence intersection compact hai.
Prove no continuous \(f:[0,1]\to\mathbb R\) can have image (a) \((0,1)\), or (b) \([0,1]\cup[2,3]\).
(a) A continuous image of the compact domain \([0,1]\) must be compact. But \((0,1)\) is not closed and hence is not compact. Therefore it cannot be the image.
(b) A continuous image of the connected interval \([0,1]\) must be connected, hence an interval in \(\mathbb R\). But \([0,1]\cup[2,3]\) contains \(1\) and \(2\) but not \(3/2\), so it is not an interval. Therefore it cannot be the image.
(a) Continuous function compactness preserve karta hai. Domain compact hai toh image compact honi chahiye, lekin open interval endpoints miss karta hai aur compact nahi. Contradiction.
(b) Domain connected interval hai. Continuous map usko gap wali image mein nahi tod sakta. Proposed image mein \(1\) aur \(2\) hain lekin beech ka \(3/2\) missing hai, so connected nahi. Aisa function exist nahi karta.
Questions 7–9 · 12 Marks
Continuity & Differentiability
Show \(f(x)=x^2\) is not uniformly continuous on \([1,\infty)\), but is uniformly continuous on \([1,2]\).
Failure on \([1,\infty)\). Set \(\varepsilon=1\). For any \(0<\delta<1\), choose \(x=1/\delta\) and \(y=1/\delta+\delta/2\). Then \(|x-y|=\delta/2<\delta\), but \[|f(y)-f(x)|=\left|\left(\frac1\delta+\frac\delta2\right)^2-\frac1{\delta^2}\right|=1+\frac{\delta^2}{4}\ge1.\] Thus uniform continuity fails.
Success on \([1,2]\). The function is continuous and the interval is compact. By Heine–Cantor, \(f\) is uniformly continuous on \([1,2]\).
Unbounded interval. Uniform continuity mein ek hi \(\delta\) poori domain ke liye kaam karna chahiye. Large x par parabola bahut steep hota hai. Inputs ko \(\delta/2\) paas rakhne ke baad bhi outputs ka difference at least \(1\) milta hai, so global \(\delta\) impossible hai.
Compact interval. \([1,2]\) compact hai aur \(x^2\) continuous hai. Heine–Cantor theorem directly uniform continuity deta hai.
Let \(f(x)=x|x|\). Show \(f\) is differentiable on \(\mathbb R\), find \(f'\), and decide whether \(f'\) is differentiable at \(0\).
For \(x>0\), \(f(x)=x^2\), so \(f'(x)=2x\). For \(x<0\), \(f(x)=-x^2\), so \(f'(x)=-2x\). At zero, \[\frac{f(h)-f(0)}h=\frac{h|h|}{h}=|h|\longrightarrow0,\] so \(f'(0)=0\). Hence \(\boxed{f'(x)=2|x|}\) for every \(x\).
However, \(\frac{f'(h)-f'(0)}h=\frac{2|h|}{h}\) equals \(2\) for \(h>0\) and \(-2\) for \(h<0\). The two-sided limit does not exist, so \(f'\) is not differentiable at \(0\).
Absolute value ki wajah se left aur right formulas alag likho. Positive side par \(x^2\), negative side par \(-x^2\). Zero par derivative definition se difference quotient \(|h|\) milta hai, jo zero ko approach karta hai. Combined answer \(f'(x)=2|x|\) hai.
Ab \(2|x|\) ka graph zero par sharp corner banata hai. Right derivative \(2\), left derivative \(-2\) hai. Dono equal nahi, isliye \(f'\) zero par differentiable nahi.
Prove \(p(x)=x^5+3x+1\) has exactly one real root.
Existence. The polynomial is continuous, \(p(-1)=-3<0\), and \(p(0)=1>0\). By the Intermediate Value Theorem, some \(c\in(-1,0)\) satisfies \(p(c)=0\).
Uniqueness. If distinct roots \(a<b\) existed, Rolle's theorem would give \(c\in(a,b)\) with \(p'(c)=0\). But \(p'(x)=5x^4+3\ge3>0\) for every real x, a contradiction. Hence exactly one real root exists.
Root exists. Polynomial continuous hai. \(-1\) par value negative aur \(0\) par positive hai, so graph ko beech mein x-axis cross karna padega. IVT se at least one root milta hai.
Root unique hai. Agar do roots hote, Rolle's theorem ke according unke beech derivative zero hota. Lekin \(5x^4+3\) hamesha positive hai. Contradiction ki wajah se second root possible nahi.
Core tools: induction, Monotone Convergence, Cauchy criterion, Archimedean property, compactness, connectedness, uniform continuity, derivative definition, IVT, and Rolle's theorem.
Revision tip: Theorem ki conditions verify karo. Sirf theorem ka naam nahi—kyun apply hota hai, woh clear sentence mein likho.