Calculating IA, IB, IC: A Parameterization Approach

[squenshl]

I'm studying for a test.
The question is:
Let A be the straight line segment from -3-4i to 4+3i. Let B be the arc of the circle |z| = 5 going anti-clockwise from -3-4i to 4+3i. Let C be the arc of the circle |z| = 5 going anti-clockwise from -3-i to 4+3i. Define:
I_A = $$\int_A$$ 1/z dz

I_B = $$\int_B$$ 1/z dz

I_C = $$\int_C$$ 1/z dz

How do I calculate I_A, I_B, I_C? I know to write I_A as an integral by parameterizing A, but how do I parameterize A, I know if I can find I_A then the other two are easy. Could I let z = e^it so dz = ie^it dt. I can't seem to get the limits integration.

 

[jackmell]

You have:

$\int_{-3-4i}^{4+3i} \frac{1}{z}dz$

Just figure out the equation of the straight line from start to end. I get simply $y=x-1$. So let $x(t)=t$ and $y(t)=t-1$ and solve:

$\int_{-3}^4 \frac{1}{x(t)+iy(t)}(dx(t)+idy(t))$

 

[HallsofIvy]

|-3-4i|= |4+ 3i|= 5 so they both lie on the circle |z|= 5 but $|-3- i|= \sqrt{10}\ne 5$ so the circle |z|= 5 does NOT go "from -3- i to 4+ 3i". Did you mean "from -3- 4i to 4+ 3i"? But then B and C are the same!? Perhaps you mean "from 4+ 3i to -3- 4i"? That is, B and C are two halves of the same circle.

Since z= x+ iy is on the circle |z|= 5 if and only if $x^2+ y^2= 5$, we can use $x= 5cos(\theta)$ and [/itex]y= 5 sin(\theta)[/itex] (which is the same as z= 5e^{i\theta}[/itex] for the circle. The only "difficulty" is getting the beginning and ending values for $\theta$. At z= 4+ 3i, $x= 5 cos(\theta)= 4$ and $5 sin(\theta)= 3$ so $(5 sin(\theta))/(5 cos(\theta))= tan(\theta)= 3/4$. For the anti-clockwise circle from 4+ 3i to -3- 4i, $\theta$ goes from $arctan(3/4)$ to $\pi+ arctan(3/4)$ and for the anti-clockwise circle from -3- 4i to 4+ 3i, $\theta$ goes from $\pi+ arctan(3/4)$ to $2\pi+ arctan(3/4)$.

 

[squenshl]

Sorry C is the arc of a circle |z| = 5 going clockwise from -3-4i to 4+3i.
How do I parameterize z
Is $$\vartheta$$ the same as for the anticlockwise arc.

 

[squenshl]

Sorry z = 5exp(i$$\vartheta$$)

 

[squenshl]

For the anti-clockwise arc I got I_B = ipi
Does that mean the clockwise arc is ipi by symmetry

 

[elibj123]

No.. 1/z is not a symmetric function.
It is also not an analytic function (in the origin), and I guess this exercise comes to teach you what it means about its line integrals.

 

[GRFrones]

1/z is analytic inside the region determined by the curves A and B and on the curves... so, $$\displaystyle\int_B f(z) dz - \displaystyle\int_A f(z) dz = 0$$ (the minus sign is to correct orientation of the curves). So, if you found B, you have A.

Joining C with B (or A, it doesn't really matter), correcting the orientations, you can find the integral using residue theorem (which is simple for f(z) = 1/z, as it's already a Laurent series, and the pole is simple (order 1)). So, if you have B, and the residue, you also have C, and the problem is solved.