WHY centre at a? Because convergence spreads symmetrically outward from one special point. That point is where convergence is guaranteed, so we measure distance ∣x−a∣ from it.
This is the Cauchy–Hadamard idea. We test absolute convergence using the Root Test on the terms bn=∣cn(x−a)n∣=∣cn∣∣x−a∣n.
Because the condition is ∣x−a∣<R — a distance condition — the solution set is automatically the symmetric interval around a. That is why it is always an interval.
For ∑3n(x−5)n, predictR before reading on.
▶ Coefficients cn=1/3n, ratio ∣cn/cn+1∣=3n+1/3n=3, so R=3, centred at 5. Geometric: both endpoints give ∑(±1)n which diverge. Interval (2,8). Did your forecast match?
What is the centre of ∑cn(x−a)n?
The point a; the series always converges there (gives c0).
Cauchy–Hadamard formula for 1/R (general, always valid)?
1/R=limsupn→∞n∣cn∣, with 1/0=∞, 1/∞=0.
Ratio-test formula for R — when is it valid?
R=limn→∞∣cn/cn+1∣, ONLY when that limit exists; otherwise use the limsup root formula.
What happens at ∣x−a∣=R?
Inconclusive — must test each endpoint separately by substitution.
What does R=0 mean?
Converges only at the centre x=a.
What does R=∞ mean?
Converges for all real x (e.g. ex series).
Why is the convergence set always an interval?
The convergence condition is a distance condition ∣x−a∣<R, symmetric about a.
Inside ∣x−a∣<R, is convergence absolute or conditional?
Absolute convergence.
R of ∑n!xn?
0 (diverges for all x=0).
R of ∑xn/n!?
∞.
Why does Cauchy–Hadamard use limsup instead of lim?
Because limsup always exists in [0,∞], so the formula gives R even when the ordinary limit fails.
Recall Feynman: explain to a 12-year-old
Imagine a campfire at point a. The closer you stand, the warmer (the series "works"). There's a magic distance R — inside it you're warm (converges), outside it you're freezing (diverges). Exactly on the circle of distance R, it's a coin-flip: you have to walk to each edge and feel for yourself whether it's warm. The campfire is the centre, the magic distance is the radius, and the warm zone (edges maybe included) is the interval of convergence.
Power series ko aise samjho jaise ek infinite-degree polynomial: ∑cn(x−a)n. Yahan a hai centre — yeh wo point hai jahan series hamesha converge karti hai (kyunki x=a daalne par sirf c0 bachta hai). Ab sawaal yeh hai: centre se kitni door tak series "kaam" karti hai? Us distance ko bolte hain radius of convergenceR.
R nikalne ka asaan tareeka hai Ratio Test: ∣an+1/an∣ ka limit lo, ismein ∣x−a∣ bahar nikal aata hai. Formula: R=lim∣cn/cn+1∣ — lekin yeh sirf tabhi valid hai jab yeh limit exist kare. Agar coefficients oscillate karein aur limit na aaye, to Ratio Test fail kar jaata hai. Tab hamesha kaam karne wala master formula hai Cauchy–Hadamard: 1/R=limsupn→∞∣cn∣1/n. limsup hamesha exist karta hai [0,∞] mein, isliye yeh formula har case mein R de deta hai.
Sabse important baat — endpoints. Jab ∣x−a∣=R, tab Ratio Test ka limit exactly 1 aata hai, jo inconclusive hota hai. Isliye dono endpoints x=a−R aur x=a+R ko alag-alag plug karke check karo (alternating series test, p-series, ya divergence test se). Tabhi pata chalega ki interval mein bracket round hoga ( ya square [.
Yeh topic isliye zaroori hai kyunki Taylor/Maclaurin series, jaise ex, sinx, ln(1+x) — sab power series hain, aur unka interval of convergence batata hai ki approximation kahan tak valid hai. Bina radius jaane tum galat jagah series use kar baithoge!