Intuition The big pictureA real integral like ∫ − ∞ ∞ d x 1 + x 2 \int_{-\infty}^{\infty}\frac{dx}{1+x^2} ∫ − ∞ ∞ 1 + x 2 d x real line , but the integrand is just the restriction of a complex function f ( z ) = 1 1 + z 2 f(z)=\frac{1}{1+z^2} f ( z ) = 1 + z 2 1 close the real line into a loop in the complex plane , integrate around the loop (easy — just sum residues at the poles trapped inside), then throw away the arc because it contributes nothing. What's left is your real integral. We trade a hard analysis problem for finding poles and computing residues — pure algebra.
Definition Isolated singularity & Laurent seriesIf f f f 0 < ∣ z − z 0 ∣ < R 0<|z-z_0|<R 0 < ∣ z − z 0 ∣ < R Laurent series
f ( z ) = ∑ n = − ∞ ∞ c n ( z − z 0 ) n . f(z)=\sum_{n=-\infty}^{\infty} c_n (z-z_0)^n. f ( z ) = ∑ n = − ∞ ∞ c n ( z − z 0 ) n . residue of f f f z 0 z_0 z 0 Res z = z 0 f = c − 1 . \boxed{\operatorname{Res}_{z=z_0} f = c_{-1}}. Res z = z 0 f = c − 1 .
WHY is c − 1 c_{-1} c − 1 Integrate the series term-by-term around a small circle ∣ z − z 0 ∣ = ρ |z-z_0|=\rho ∣ z − z 0 ∣ = ρ z = z 0 + ρ e i θ z=z_0+\rho e^{i\theta} z = z 0 + ρ e i θ ∮ ( z − z 0 ) n d z = ∫ 0 2 π ρ n e i n θ i ρ e i θ d θ = i ρ n + 1 ∫ 0 2 π e i ( n + 1 ) θ d θ . \oint (z-z_0)^n\,dz=\int_0^{2\pi}\!\rho^n e^{in\theta}\,i\rho e^{i\theta}\,d\theta=i\rho^{n+1}\!\int_0^{2\pi}\! e^{i(n+1)\theta}d\theta. ∮ ( z − z 0 ) n d z = ∫ 0 2 π ρ n e in θ i ρ e i θ d θ = i ρ n + 1 ∫ 0 2 π e i ( n + 1 ) θ d θ . θ \theta θ 2 π 2\pi 2 π n + 1 = 0 n+1=0 n + 1 = 0 n = − 1 n=-1 n = − 1 zero otherwise (a full sinusoid averages to 0). So
∮ f d z = 2 π i c − 1 = 2 π i Res z 0 f . \oint f\,dz = 2\pi i\, c_{-1}=2\pi i\operatorname{Res}_{z_0}f. ∮ f d z = 2 π i c − 1 = 2 π i Res z 0 f . c − 1 c_{-1} c − 1 that's why we care about it.
Goal: I = ∫ − ∞ ∞ R ( x ) d x \displaystyle I=\int_{-\infty}^{\infty} R(x)\,dx I = ∫ − ∞ ∞ R ( x ) d x R = P Q R=\frac{P}{Q} R = Q P deg Q ≥ deg P + 2 \deg Q\ge \deg P+2 deg Q ≥ deg P + 2
Intuition WHAT we do, HOW, WHYWHAT: Close the contour with a big semicircle Γ ρ \Gamma_\rho Γ ρ ρ \rho ρ upper half-plane (UHP).
HOW: ∮ C ρ = ∫ − ρ ρ R d x + ∫ Γ ρ R d z = 2 π i ∑ ( Res in UHP ) \oint_{C_\rho}=\int_{-\rho}^{\rho}R\,dx + \int_{\Gamma_\rho}R\,dz = 2\pi i\sum(\text{Res in UHP}) ∮ C ρ = ∫ − ρ ρ R d x + ∫ Γ ρ R d z = 2 π i ∑ ( Res in UHP ) WHY the arc dies: on Γ ρ \Gamma_\rho Γ ρ ∣ R ∣ ∼ c ρ 2 |R|\sim \frac{c}{\rho^{2}} ∣ R ∣ ∼ ρ 2 c π ρ \pi\rho π ρ ∣ ∫ Γ ρ ∣ ≲ c ρ 2 ⋅ π ρ → 0 \big|\int_{\Gamma_\rho}\big|\lesssim \frac{c}{\rho^{2}}\cdot\pi\rho\to 0 ∫ Γ ρ ≲ ρ 2 c ⋅ π ρ → 0 ≥ 2 \ge 2 ≥ 2 exactly what's needed.
Taking ρ → ∞ \rho\to\infty ρ → ∞ ∫ − ∞ ∞ R ( x ) d x = 2 π i ∑ poles in UHP Res R . \boxed{\int_{-\infty}^{\infty}R(x)\,dx = 2\pi i\sum_{\text{poles in UHP}}\operatorname{Res} R.} ∫ − ∞ ∞ R ( x ) d x = 2 π i poles in UHP ∑ Res R .
∫ − ∞ ∞ d x 1 + x 2 \displaystyle\int_{-\infty}^{\infty}\frac{dx}{1+x^2} ∫ − ∞ ∞ 1 + x 2 d x Poles: z 2 + 1 = 0 ⇒ z = ± i z^2+1=0\Rightarrow z=\pm i z 2 + 1 = 0 ⇒ z = ± i z = i z=i z = i Why only i i i Because we closed upward.
Residue (simple pole, use g / h ′ g/h' g / h ′ g = 1 g=1 g = 1 h = 1 + z 2 h=1+z^2 h = 1 + z 2 h ′ = 2 z h'=2z h ′ = 2 z Res i = 1 2 i \operatorname{Res}_{i}=\frac{1}{2i} Res i = 2 i 1 Why g / h ′ g/h' g / h ′ It avoids factoring; quick for simple poles.
Assemble: I = 2 π i ⋅ 1 2 i = π I=2\pi i\cdot\frac{1}{2i}=\pi I = 2 π i ⋅ 2 i 1 = π arctan x ∣ − ∞ ∞ = π \arctan x\big|_{-\infty}^{\infty}=\pi arctan x − ∞ ∞ = π
∫ − ∞ ∞ d x ( x 2 + 1 ) 2 \displaystyle\int_{-\infty}^{\infty}\frac{dx}{(x^2+1)^2} ∫ − ∞ ∞ ( x 2 + 1 ) 2 d x Pole in UHP: z = i z=i z = i m = 2 m=2 m = 2 Why order 2? The factor ( z 2 + 1 ) 2 = ( z − i ) 2 ( z + i ) 2 (z^2+1)^2=(z-i)^2(z+i)^2 ( z 2 + 1 ) 2 = ( z − i ) 2 ( z + i ) 2 Residue: Res i = 1 1 ! d d z [ ( z − i ) 2 1 ( z − i ) 2 ( z + i ) 2 ] z = i = d d z ( z + i ) − 2 ∣ i = − 2 ( z + i ) 3 ∣ i . \operatorname{Res}_i=\frac{1}{1!}\frac{d}{dz}\Big[(z-i)^2\frac{1}{(z-i)^2(z+i)^2}\Big]_{z=i}=\frac{d}{dz}(z+i)^{-2}\Big|_i=\frac{-2}{(z+i)^3}\Big|_i. Res i = 1 ! 1 d z d [ ( z − i ) 2 ( z − i ) 2 ( z + i ) 2 1 ] z = i = d z d ( z + i ) − 2 i = ( z + i ) 3 − 2 i . z = i z=i z = i ( 2 i ) 3 = 8 i 3 = − 8 i (2i)^3=8i^3=-8i ( 2 i ) 3 = 8 i 3 = − 8 i Res i = − 2 − 8 i = 1 4 i \operatorname{Res}_i=\frac{-2}{-8i}=\frac{1}{4i} Res i = − 8 i − 2 = 4 i 1 Why differentiate once? m − 1 = 1 m-1=1 m − 1 = 1 Answer: I = 2 π i ⋅ 1 4 i = π 2 I=2\pi i\cdot\frac{1}{4i}=\frac{\pi}{2} I = 2 π i ⋅ 4 i 1 = 2 π
For ∫ R ( x ) cos ( a x ) d x \int R(x)\cos(ax)\,dx ∫ R ( x ) cos ( a x ) d x ∫ R ( x ) sin ( a x ) d x \int R(x)\sin(ax)\,dx ∫ R ( x ) sin ( a x ) d x a > 0 a>0 a > 0 e i a z e^{iaz} e ia z
∫ − ∞ ∞ cos x x 2 + 1 d x \displaystyle\int_{-\infty}^{\infty}\frac{\cos x}{x^2+1}\,dx ∫ − ∞ ∞ x 2 + 1 cos x d x f ( z ) = e i z z 2 + 1 f(z)=\frac{e^{iz}}{z^2+1} f ( z ) = z 2 + 1 e i z a = 1 a=1 a = 1 Why e i z e^{iz} e i z cos z \cos z cos z ∣ cos z ∣ |\cos z| ∣ cos z ∣ e i z e^{iz} e i z Pole in UHP: z = i z=i z = i Res i = e i ⋅ i 2 i = e − 1 2 i \operatorname{Res}_i=\frac{e^{i\cdot i}}{2i}=\frac{e^{-1}}{2i} Res i = 2 i e i ⋅ i = 2 i e − 1 ∫ − ∞ ∞ e i x x 2 + 1 d x = 2 π i ⋅ e − 1 2 i = π e \int_{-\infty}^\infty\frac{e^{ix}}{x^2+1}dx=2\pi i\cdot\frac{e^{-1}}{2i}=\frac{\pi}{e} ∫ − ∞ ∞ x 2 + 1 e i x d x = 2 π i ⋅ 2 i e − 1 = e π real part : ∫ cos x x 2 + 1 d x = π e \int\frac{\cos x}{x^2+1}dx=\frac{\pi}{e} ∫ x 2 + 1 c o s x d x = e π 0 0 0
Intuition Turn the circle into THE circleAn integral ∫ 0 2 π F ( cos θ , sin θ ) d θ \int_0^{2\pi}F(\cos\theta,\sin\theta)\,d\theta ∫ 0 2 π F ( cos θ , sin θ ) d θ already a loop. Substitute z = e i θ z=e^{i\theta} z = e i θ unit circle ∣ z ∣ = 1 |z|=1 ∣ z ∣ = 1
Using z = e i θ z=e^{i\theta} z = e i θ cos θ = z + z − 1 2 \cos\theta=\frac{z+z^{-1}}{2} cos θ = 2 z + z − 1 sin θ = z − z − 1 2 i \sin\theta=\frac{z-z^{-1}}{2i} sin θ = 2 i z − z − 1 d z = i e i θ d θ ⇒ d θ = d z i z dz=ie^{i\theta}d\theta\Rightarrow d\theta=\frac{dz}{iz} d z = i e i θ d θ ⇒ d θ = i z d z
∫ 0 2 π d θ 2 + cos θ \displaystyle\int_0^{2\pi}\frac{d\theta}{2+\cos\theta} ∫ 0 2 π 2 + cos θ d θ = ∮ ∣ z ∣ = 1 1 2 + z + z − 1 2 d z i z = ∮ 2 i ( z 2 + 4 z + 1 ) d z . =\oint_{|z|=1}\frac{1}{2+\frac{z+z^{-1}}{2}}\frac{dz}{iz}=\oint \frac{2}{i(z^2+4z+1)}\,dz. = ∮ ∣ z ∣ = 1 2 + 2 z + z − 1 1 i z d z = ∮ i ( z 2 + 4 z + 1 ) 2 d z . Poles: z 2 + 4 z + 1 = 0 ⇒ z = − 2 ± 3 z^2+4z+1=0\Rightarrow z=-2\pm\sqrt3 z 2 + 4 z + 1 = 0 ⇒ z = − 2 ± 3 ∣ z ∣ < 1 |z|<1 ∣ z ∣ < 1 − 2 + 3 ≈ − 0.27 -2+\sqrt3\approx-0.27 − 2 + 3 ≈ − 0.27 − 2 − 3 ≈ − 3.73 -2-\sqrt3\approx-3.73 − 2 − 3 ≈ − 3.73 Residue at z 0 = − 2 + 3 z_0=-2+\sqrt3 z 0 = − 2 + 3 2 i ⋅ 1 2 z 0 + 4 = 2 i ⋅ 1 2 3 = 1 i 3 . \frac{2}{i}\cdot\frac{1}{2z_0+4}=\frac{2}{i}\cdot\frac{1}{2\sqrt3}=\frac{1}{i\sqrt3}. i 2 ⋅ 2 z 0 + 4 1 = i 2 ⋅ 2 3 1 = i 3 1 . Answer: 2 π i ⋅ 1 i 3 = 2 π 3 2\pi i\cdot\frac{1}{i\sqrt3}=\frac{2\pi}{\sqrt3} 2 π i ⋅ i 3 1 = 3 2 π
Common mistake "I'll just sum residues atall poles."
Why it feels right: the residue theorem says "sum the residues." The fix: only the poles enclosed by your contour count. With a UHP semicircle, lower-half poles (z = − i z=-i z = − i outside — ignore them. Picking the wrong half flips signs.
cos z \cos z cos z ∫ R ( x ) cos ( a x ) d x \int R(x)\cos(ax)dx ∫ R ( x ) cos ( a x ) d x Why it feels right: you want a cosine, so write a cosine. The fix: cos z = e i z + e − i z 2 \cos z=\frac{e^{iz}+e^{-iz}}{2} cos z = 2 e i z + e − i z e − i z e^{-iz} e − i z explodes in the UHP. Use e i a z e^{iaz} e ia z
Common mistake Forgetting the degree/decay condition.Why it feels right: the formula "just works." The fix: the arc vanishes only if deg Q ≥ deg P + 2 \deg Q\ge \deg P+2 deg Q ≥ deg P + 2
Common mistake Dropping the factor ofi i i Why it feels right: d θ ↔ d z d\theta\leftrightarrow dz d θ ↔ d z The fix: d θ = d z i z d\theta=\frac{dz}{iz} d θ = i z d z 1 i z \frac1{iz} i z 1 adds another pole at z = 0 z=0 z = 0 . Check it!
Recall Feynman: explain to a 12-year-oldImagine a real number line — a long straight road where your integral runs. Now lift the road up into a 2-D map (the complex plane) and bend it into a giant loop, like catching a string and pulling both ends together. Inside the loop there might be a few "magic spikes" (poles). There's a rule: the loop's total only depends on how strong those spikes are — a number called the residue of each spike. The far-away part of the loop (the big arc) is so weak it adds nothing. So the whole hard road-integral equals just 2 π 2\pi 2 π i i i
"CLOSE UP, the ARC DROPS, sum what's TRAPPED."
CLOSE UP → semicircle in the U pper half-plane.ARC DROPS → check decay (deg gap ≥ 2 \ge2 ≥ 2 TRAPPED → only enclosed poles; multiply sum of residues by 2 π i 2\pi i 2 π i
Cauchy's Integral Theorem & Formula (residue theorem is their generalization)Laurent Series & Classification of Singularities (defines c − 1 c_{-1} c − 1 Jordan's Lemma (controls oscillatory arcs)Contour Integration & Branch Cuts (keyhole contours for ∫ 0 ∞ x s − 1 / ( 1 + x ) d x \int_0^\infty x^{s-1}/(1+x)\,dx ∫ 0 ∞ x s − 1 / ( 1 + x ) d x Principal Value Integrals (indented contours around real poles)Argument Principle & Rouché's Theorem (counting zeros via residues)
What is the residue of f f f z 0 z_0 z 0 The coefficient
c − 1 c_{-1} c − 1 of
( z − z 0 ) − 1 (z-z_0)^{-1} ( z − z 0 ) − 1 .
Why does only c − 1 c_{-1} c − 1 z 0 z_0 z 0 ∮ ( z − z 0 ) n d z = 2 π i \oint(z-z_0)^n dz=2\pi i ∮ ( z − z 0 ) n d z = 2 π i only for
n = − 1 n=-1 n = − 1 ; all other powers integrate to
0 0 0 over a full circle.
State the residue theorem. ∮ C f d z = 2 π i ∑ enclosed poles Res f \oint_C f\,dz = 2\pi i\sum_{\text{enclosed poles}}\operatorname{Res} f ∮ C f d z = 2 π i ∑ enclosed poles Res f for
C C C positively oriented.
Residue formula for a simple pole of g / h g/h g / h h ( z 0 ) = 0 h(z_0)=0 h ( z 0 ) = 0 Res z 0 = g ( z 0 ) / h ′ ( z 0 ) \operatorname{Res}_{z_0}=g(z_0)/h'(z_0) Res z 0 = g ( z 0 ) / h ′ ( z 0 ) .
Residue formula for a pole of order m m m 1 ( m − 1 ) ! lim z → z 0 d m − 1 d z m − 1 [ ( z − z 0 ) m f ] \frac{1}{(m-1)!}\lim_{z\to z_0}\frac{d^{m-1}}{dz^{m-1}}[(z-z_0)^m f] ( m − 1 )! 1 lim z → z 0 d z m − 1 d m − 1 [( z − z 0 ) m f ] .
Condition for the semicircular arc to vanish for a rational integrand. deg ( denominator ) ≥ deg ( numerator ) + 2 \deg(\text{denominator})\ge \deg(\text{numerator})+2 deg ( denominator ) ≥ deg ( numerator ) + 2 .
Why use e i a z e^{iaz} e ia z cos ( a z ) \cos(az) cos ( a z ) ∣ e i a z ∣ = e − a I m z |e^{iaz}|=e^{-a\,\mathrm{Im}\,z} ∣ e ia z ∣ = e − a Im z decays in the UHP;
cos z \cos z cos z grows there. Take Re/Im at the end.
What does Jordan's lemma let you weaken compared to the rational case? The arc vanishes needing only
R ( z ) → 0 R(z)\to0 R ( z ) → 0 (degree gap
≥ 1 \ge1 ≥ 1 ), because the exponential supplies decay.
Substitution to convert ∫ 0 2 π F ( cos θ , sin θ ) d θ \int_0^{2\pi}F(\cos\theta,\sin\theta)d\theta ∫ 0 2 π F ( cos θ , sin θ ) d θ z = e i θ z=e^{i\theta} z = e i θ ,
cos θ = z + z − 1 2 \cos\theta=\frac{z+z^{-1}}{2} cos θ = 2 z + z − 1 ,
sin θ = z − z − 1 2 i \sin\theta=\frac{z-z^{-1}}{2i} sin θ = 2 i z − z − 1 ,
d θ = d z i z d\theta=\frac{dz}{iz} d θ = i z d z , over
∣ z ∣ = 1 |z|=1 ∣ z ∣ = 1 .
Value of ∫ − ∞ ∞ d x 1 + x 2 \int_{-\infty}^\infty \frac{dx}{1+x^2} ∫ − ∞ ∞ 1 + x 2 d x 2 π i ⋅ 1 2 i = π 2\pi i\cdot\frac{1}{2i}=\pi 2 π i ⋅ 2 i 1 = π (pole at
z = i z=i z = i ).
Value of ∫ − ∞ ∞ cos x x 2 + 1 d x \int_{-\infty}^\infty \frac{\cos x}{x^2+1}dx ∫ − ∞ ∞ x 2 + 1 c o s x d x Which poles do you sum when closing in the upper half-plane? Only those with positive imaginary part (inside the contour).
loop integral equals 2 pi i sum res
Real integral on real line
Arc vanishes as rho to infinity
Integral equals 2 pi i sum UHP residues
Intuition Hinglish mein samjhoDekho, idea bahut simple hai. Tumhe ek real integral solve karna hai jaise ∫ − ∞ ∞ d x 1 + x 2 \int_{-\infty}^{\infty}\frac{dx}{1+x^2} ∫ − ∞ ∞ 1 + x 2 d x x x x z z z loop bana dete hain upper half-plane mein. Complex analysis ka jaadu yeh hai: aise band loop ka integral sirf andar trapped poles ke "residue" pe depend karta hai — 2 π i × ( sum of residues ) 2\pi i \times (\text{sum of residues}) 2 π i × ( sum of residues )
Trick ka second part: jo bada arc (semicircle ka curved hissa) hai, woh radius bada karne par zero ho jaata hai — provided denominator ki degree numerator se kam se kam 2 zyada ho. Toh jab ρ → ∞ \rho\to\infty ρ → ∞ π \pi π z = i z=i z = i 1 2 i \frac{1}{2i} 2 i 1 2 π i 2\pi i 2 π i
Oscillatory waale, jaise cos x x 2 + 1 \frac{\cos x}{x^2+1} x 2 + 1 c o s x cos z \cos z cos z e i z e^{iz} e i z e i z e^{iz} e i z e − I m z e^{-\mathrm{Im}\,z} e − Im z cos z \cos z cos z 0 0 0 2 π 2\pi 2 π z = e i θ z=e^{i\theta} z = e i θ d θ = d z i z d\theta=\frac{dz}{iz} d θ = i z d z
Sabse important sawaal exam mein: kaunse pole count karne hain? Sirf woh jo tumhare loop ke andar hain. UHP semicircle ho to sirf positive imaginary part waale poles. Bas yeh discipline rakho — half choose karo, arc decay check karo, trapped residues add karo — aur tough-looking integrals 2 line mein khatam.
