diff options
Diffstat (limited to 'kang/hw4.tex')
-rw-r--r-- | kang/hw4.tex | 163 |
1 files changed, 163 insertions, 0 deletions
diff --git a/kang/hw4.tex b/kang/hw4.tex new file mode 100644 index 0000000..d4fb394 --- /dev/null +++ b/kang/hw4.tex @@ -0,0 +1,163 @@ +\def\Re{\mathop{\rm Re}\nolimits} +\def\Im{\mathop{\rm Im}\nolimits} +\let\rule\hrule +\def\hrule{\medskip\rule\medskip} + +%page 70 +\noindent{\bf 1.} + +\noindent{\it (a)} + +If $f(z) = \overline z = x - yi,$ $U_x = 1$ and $V_y = -1.$ $U_x\neq V_y,$ +so because the Cauchy-Riemann equations are dissatisfied, the derivative +does not exist. + +\noindent{\it (b)} + +If $f(z) = z-\overline z = x + yi - (x - yi) = 2yi,$ $U_x = 0,$ and $V_y += 2i.$ Again, $U_x\neq V_y.$ + +\noindent{\it (c)} + +If $f(z) = 2x+ixy^2,$ + +$U_x = 2.$ $V_y = 2ix.$ +$U_x = V_y \to 2 = 2ix \to x = 1/i = -i.$ $x\in R,$ so this is false, +and the function is not analytic. + +\noindent{\it (d)} + +If $f(z) = e^xe^{-iy},$ + +$$U_x = e^xe^{-iy}.$$ +$$V_y = -ie^xe^{-iy} \neq U_x.$$ + +\noindent{\bf 3.} +% only info from secs 21 and 23 + +\noindent{\it (a)} + +$f(z) = 1/z = \overline z/|z|^2 = (a-bi)/(a^2+b^2).$ + +The partial derivatives are $U_x = (y^2-x^2)/(x^2+y^2)^2,$ +$V_y = (y^2-x^2)/(x^2+y^2)^2,$ $U_y = (-2xy)/(x^2+y^2)^2,$ and $V_x = +(2xy)/(x^2+y^2)^2.$ These satisfy the Cauchy-Riemann equations except +where they are undefined, $x^2+y^2 = 0 \to z = 0.$ Using the power rule +on $f(z) = z^{-1},$ $f'(z) = -z^{-2}.$ + +\noindent{\it (b)} + +$f(z) = x^2+iy^2.$ + +The partial derivatives are $U_x = 2x,$ $U_y = 0,$ $V_y = 2y,$ and $V_x += 0.$ $0 = 0$ at all points, but $2y = 2x$ only for $x = y$ or +$z = x+ix.$ + +Its derivative is, from these partial derivatives, $2x.$ + +\noindent{\it (c)} + +$f(z) = z\Im z = (x+iy)y = xy + iy^2.$ + +$U_x = y = V_y = 2y \to y = 0.$ $U_y = x = -V_x = 0 \to x = 0.$ +Therefore, $z = 0 + i0 = 0.$ From these partial derivatives, its +derivative on this domain is $0.$ + +\hrule +%page 76 +\noindent{\bf 4.} + +\noindent{\it (a)} + +$$f(z) = {2z+1\over z(z^2+1)}$$ + +$f(z)$ is not analytic on $z=0$ or $z^2+1=0 \to z = \pm i.$ It is, +however, analytic on the remainder of the plane because, using the chain +rule and the product rule, its derivative can be constructed from its +component polynomials (which are entire functions). + +\noindent{\it (b)} + +$$f(z) = {z^3+i\over z^2-3z+2}$$ + +This is not analytic on $z^2-3z+2 = 0 = (z-1)(z-2) \to z = 1,2.$ It is +analytic on the remainder of the plane for the same reason as (a). + +\noindent{\it (c)} + +$$f(z) = {z^2+1\over (z+2)(z^2+2z+2)}.$$ + +This is not analytic on $z+2 = 0 \to z = -2$ or $z^2+2z+2 = 0 = +(z+(1-i))(z+(1+i)) \to z = 1\pm i.$ It is analytic for the same reason +as (a). + +\hrule +%page 79 +\noindent{\bf 4.} + +See addendum. + +\hrule +%page 89 +\noindent{\bf 2.} + +$$f(z) = 2z^2 - 3 - ze^z + e^{-z}.$$ + +If the component functions $f$ and $g$ are analytic on the plane (have a +defined derivative), $f+g$ is also analytic on the plane. +A similar rule exists for the product of entire functions. +$2z^2$ follows the power rule, and has a derivative over the entire +plane, and $-3$ follows the constant rule, having a derivative of 0. +$-z$ is entire, also from the power rule, and $e^z$ has been shown to +have a derivative in chapter 3. Using the product rule, $-ze^z$ is +entire. +$e^{-z}$ is a composition of $e^z$ and $-z,$ which each are analytic +over the plane, so their composition is also analytic over the plane. + +This means $f(z)$ is entire. + +\noindent{\bf 8.} + +\noindent{\it (a)} + +$e^z = -2.$ $e^x = |z| = |-2| = 2 \to x = \ln 2,$ and +$\arg z = \pi + 2n\pi,$ so $z = \ln2 + i(\pi+2n\pi).$ + +\noindent{\it (b)} + +$e^z = 1 + i.$ $e^x = |z| = |1+i| = \sqrt 2 \to x = (1/2)\ln2,$ and +$\arg z = \pi/4 + 2n\pi.$ Therefore, $z = (1/2)\ln2 + i(\pi/4+2n\pi).$ + +\noindent{\it (c)} + +If $\exp(2z-1) = 1,$ $\exp(2z)e^{-1} = 1 \to e^{2z} = e.$ This gives +$e^{2x} = |e| = e \to 2x = 1 \to x = 1/2,$ and $\arg(2z) = 2n\pi \to +\arg z = n\pi,$ so $z = 1/2 + in\pi.$ + +\hrule +%page 95 +\noindent{\bf 2.} + +\noindent{\it (a)} + +$$\log e = 1 + 2n\pi i$$ + +$\ln e = 1,$ and $\arg e = 2n\pi.$ This proves the above identity (with +$n = 0,\pm1,\pm2,\ldots$) + +\noindent{\it (b)} + +$$\log i = \left(2n + {1\over2}\right)\pi i$$ + +$\ln|i| = \ln1 = 0,$ and $\arg i = \pi/2 + 2n\pi.$ This makes $\log i = +(2n+1/2)\pi i.$ + +\noindent{\it (c)} + +$$\log(-1+\sqrt3i) = \ln2 + 2\left(n + {1\over3}\right)\pi i$$ + +$\ln|-1+\sqrt3i| = \ln2,$ and $\arg(-1+\sqrt3i) = 2\pi/3 + 2n\pi,$ which +corresponds to the given identity ($\log(-1+\sqrt3i) = (2n\pi+2\pi/3)i + +\ln2.$) + +\bye |