Yeh page tumhara self-test hai. Har problem L1 (Recognition) se lekar L5 (Mastery) tak graded hai. Problem padho, solution kholne se pehle try karo, phir check karo. Sab kuch parent note se connect hota hai.
WHAT:f(x)=x2+11 lo. WHY legal (teenon hypotheses): positive ✓, continuous ✓; f′(x)=(x2+1)2−2x<0 for x≥1 to decreasing ✓.
∫1∞x2+1dx=limb→∞[arctanx]1b=limb→∞(arctanb−arctan1)=2π−4π=4π.WHY arctan?x2+11 ka antiderivative arctanx hai — yeh woh ek integral hai jiska value ek angle hai, aur arctanb→2π as b→∞. Integral finite hai (π/4≈0.785) ⇒ series converges.
Recall Solution 2.2
WHAT:f(x)=xe−x2, [1,∞) par positive hai; f′(x)=e−x2(1−2x2)<0 for x≥1 ⇒ decreasing ✓.
WHY substitution u=x2: aage wala factor x exactly 21du hai jahan du=2xdx — yeh pure integrand ko e−u mein collapse kar deta hai.
∫1∞xe−x2dx=u=x221∫1∞e−udu=21[−e−u]1∞=21e−1=2e1.
Finite (≈0.184) ⇒ series converges.
KAHAN decrease karta hai:f′(x)=x21−lnx. Yeh <0 tab hai jab lnx>1, yaani x>e≈2.718. To fx=e tak increase karta hai, phir decrease karta hai.
Figure s02 mein kya dekhna hai: blue curve y=lnx/xx=1 se climb karta hai, yellow dashed line x=e par ek single peak (red dot) par pahunchta hai, phir hamesha ke liye girta hai. Peak ke left mein yeh rise karta hai (green label), isliye integral-test rectangle comparison wahan ulti direction mein jaata; peak ke right mein yeh fall karta hai (red label), jahaan test legal hai.
WHY yeh theek hai: integral test ko sirf eventually decreasing f chahiye — N=3 se start karo (tail hi convergence decide karta hai; shuru ke finite terms fate nahi badal sakte).
∫3∞xlnxdx=u=lnx∫ln3∞udu=limb→∞2u2ln3b=∞.WHY u=lnx:du=dx/xx1 ko absorb kar leta hai, sirf pure u bachta hai, jo bina bound ke grow karta hai. Integral diverge karta hai ⇒ series diverges.
Recall Solution 3.2
WHAT:f(x)=x(lnx)p1 on [2,∞). WHY positive:x≥2 ke liye, lnx>0 to denominator ke dono factors positive hain ✓. WHY continuous: yeh continuous functions ka quotient hai jiska denominator [2,∞) par kabhi zero nahi hota ✓. WHY decreasing:g(x)=x(lnx)p (denominator) likho; phir
g′(x)=(lnx)p+x⋅p(lnx)p−1⋅x1=(lnx)p−1(lnx+p)>0for x≥2,
kyunki lnx>0 aur lnx+p>0. Ek growing positive denominator ka matlab hai f=1/g decreasing hai ✓.
WHY u=lnx:du=dx/x integrand ko u mein pure p-integral bana deta hai.
∫2∞x(lnx)pdx=u=lnx∫ln2∞updu.
Ab yeh u mein ek p-integral hai, jiska convergence hum jaante hain:
p>1: ∫u−pduconverges ⇒ series converges.
p≤1: diverges ⇒ series diverges.
To answer wohi same razor cutoff hai jaise p-series: convergence ⟺p>1. (p=1 par hum parent ke Example 3 par wapas aate hain, jo diverge karta hai.)
WHY integrate nahi karte:x3+11 ka koi clean antiderivative nahi hai. Iski jagah ek p-series se compare karo jiska fate hum integral test se jaante hain.
Sabhi n≥1 ke liye: n3+1>n3=n3/2, isliye
n3+11<n3/21.WHY yeh bound help karta hai:∑n3/21 ek p-series hai jisme p=23>1, jise integral test converges prove karta hai. Ek convergent series ke neeche chhota positive series bhi converge karta hai (Comparison test). ⇒ converges.
(Cross-check Limit comparison test se: lim1/n3/21/n3+1=1, ek finite nonzero limit, isliye same fate.)
Recall Solution 4.2
Trapping inequality yaad karo jisme SN=∑n=1Nf(n)N-term partial sum hai:
∫1N+1f≤SN≤f(1)+∫1Nf.
Yahan f(x)=x−2, f(1)=1, aur ∫1∞x−2dx=1 (parent Example 1). N→∞ lo to SN→S:
=1∫1∞x2dx≤S≤=1f(1)+=1∫1∞x2dx.
To 1≤S≤2. WHY yahi test ka point hai: yeh sum ko evaluate kiye bina bound karta hai. True value π2/6≈1.645 hai, jo wakai [1,2] mein aata hai — kabhi bhi kisi integral ke equal nahi (yeh classic mistake hai). Tighten karne ke liye: tail∑n=2∞ ko ∫1∞x−2=1 se bound karo, jo S≤1+1=2 aur S≥1+∫2∞x−2=1+21=1.5 deta hai.
Idea (same rectangle geometry jaise page ke top par, tail par shift ki gayi).[n,n+1] par, decreasing f deta hai
f(n+1)≤∫nn+1f≤f(n).
Figure s03 mein kya dekhna hai: blue curve y=1/x2 tail x≥3 par hai. Red dashed rectangles left-endpoint height f(n) use karte hain aur curve ke upar nikalte hain (unka sum overshoot karta hai ⇒ upper bound); green rectangles right-endpoint height f(n+1) use karte hain aur curve ke neeche baithe hain (unka sum undershoot karta hai ⇒ lower bound). Remainder RN dono shaded areas ke beech trap hai.
RN ke liye upper bound:left inequality f(n+1)≤∫nn+1f use karke, n=N,N+1,… sum karo:
RN=∑n=N+1∞f(n)=∑k=N∞f(k+1)≤∑k=N∞∫kk+1f=∫N∞f.RN ke liye lower bound:right inequality ∫nn+1f≤f(n) use karke, n=N+1,N+2,… sum karo:
∫N+1∞f=∑n=N+1∞∫nn+1f≤∑n=N+1∞f(n)=RN.
Combine karo: ∫N+1∞f≤RN≤∫N∞f. ■
∑1/n2 par apply karo, N=3:∫N∞x−2dx=N1, isliye
41≤R3≤31,i.e. 0.25≤R3≤0.333…
(True remainder R3=π2/6−1−41−91≈0.2838 — band ke andar hai.)
Recall Solution 5.2
Partial sum S3=1+41+91=3649≈1.3611. Problem 5.1 se, 41≤R3≤31. Isliye
S3+41≤S≤S3+31,1.6111≤S≤1.6944.Midpoint estimate:S≈21.6111+1.6944=1.6528, maximum error ke saath 21.6944−1.6111≈0.0417.
True value π2/6≈1.6449[1.6111,1.6944] ke andar hai aur hamare midpoint se 0.008 ke andar — Problem 4.2 ke crude [1,2] se kaafi better. WHY better hai: humne kuch exact terms compute kiye aur sirf tail ko integrals se estimate kiya, jahaan rectangle error chota hota hai.