Respuesta :

Looks like you're looking for the extrema of [tex]f(x,y,z)=yz+xy[/tex] subject to [tex]xy=1[/tex] and [tex]y^2+z^2=1[/tex]. The Lagrangian is

[tex]L(x,y,z,\lambda,\mu)=yz+xy+\lambda(xy-1)+\mu(y^2+z^2-1)[/tex]

Look for any critical points:

[tex]L_x=y+\lambda y=y(1+\lambda)=0\implies y=0\text{ or }\lambda=-1[/tex]

[tex]L_y=z+x+\lambda x+2\mu y=0[/tex]

[tex]L_z=y+2\mu z=0[/tex]

[tex]L_\lambda=xy-1=0\implies xy=1[/tex]

[tex]L_\mu=y^2+z^2-1=0\implies y^2+z^2=1[/tex]

  • If [tex]y=0[/tex], then

[tex]L_z=0\implies2\mu z=0\implies z=0[/tex] (we don't want [tex]\lambda=\mu=0[/tex])

but then [tex]y^2+z^2=0\neq1[/tex] so we omit this case.

  • If [tex]\lambda=-1[/tex], then

[tex]L_y=0\implies z+2\mu y=0[/tex]

[tex]L_\mu=0\implies(1+4\mu^2)y^2-1=0\implies y=\pm\dfrac1{\sqrt{1+4\mu^2}}[/tex]

[tex]L_z=0\implies z=\mp\dfrac1{2\mu\sqrt{1+4\mu^2}}[/tex]

[tex]L_\mu=0\implies\dfrac1{1+4\mu^2}+\dfrac1{4\mu^2(1+4\mu^2)}=1\implies\mu=\pm\dfrac12[/tex]

[tex]L_\lambda=0\implies x=\pm\sqrt{1+4\mu^2}[/tex]

Now,

  • if [tex]\mu=\dfrac12[/tex], we have two critical points at [tex]\left(\sqrt2,\dfrac1{\sqrt2},-\dfrac1{\sqrt2}\right)[/tex] and [tex]\left(-\sqrt2,-\dfrac1{\sqrt2},\dfrac1{\sqrt2}\right)[/tex];
  • if [tex]\mu=-\dfrac12[/tex], we have two additional critical points [tex]\left(\sqrt2,\dfrac1{\sqrt2},\dfrac1{\sqrt2}\right)[/tex] and [tex]\left(-\sqrt2,-\dfrac1{\sqrt2},-\dfrac1{\sqrt2}\right)[/tex]

The first two critical points give a minimum value of [tex]f(x,y,z)=\dfrac12[/tex], and the other two give a maximum value of [tex]f(x,y,z)=\dfrac32[/tex].