The distance between an arbitrary point on the surface and the origin is
![d(x,y,z)=\sqrt{x^2+y^2+z^2}](https://tex.z-dn.net/?f=d%28x%2Cy%2Cz%29%3D%5Csqrt%7Bx%5E2%2By%5E2%2Bz%5E2%7D)
Recall that for differentiable functions
![g(x)](https://tex.z-dn.net/?f=g%28x%29)
and
![h(x)](https://tex.z-dn.net/?f=h%28x%29)
, the composition
![g(h(x))](https://tex.z-dn.net/?f=g%28h%28x%29%29)
attains extrema at the same points that
![h(x)](https://tex.z-dn.net/?f=h%28x%29)
does, so we can consider an augmented distance function
![D(x,y,z)=x^2+y^2+z^2](https://tex.z-dn.net/?f=D%28x%2Cy%2Cz%29%3Dx%5E2%2By%5E2%2Bz%5E2)
The Lagrangian would then be
![L(x,y,z,\lambda)=x^2+y^2+z^2+\lambda(y^2-64-xz)](https://tex.z-dn.net/?f=L%28x%2Cy%2Cz%2C%5Clambda%29%3Dx%5E2%2By%5E2%2Bz%5E2%2B%5Clambda%28y%5E2-64-xz%29)
We have partial derivatives
![\begin{cases}L_x=2x-\lambda z\\L_y=2y+2y\lambda\\L_z=2z-\lambda x\\L_\lambda=y^2-64-xz\end{cases}](https://tex.z-dn.net/?f=%5Cbegin%7Bcases%7DL_x%3D2x-%5Clambda%20z%5C%5CL_y%3D2y%2B2y%5Clambda%5C%5CL_z%3D2z-%5Clambda%20x%5C%5CL_%5Clambda%3Dy%5E2-64-xz%5Cend%7Bcases%7D)
Set each partial derivative to 0 and solve the system to find the critical points.
From the second equation it follows that either
![y=0](https://tex.z-dn.net/?f=y%3D0)
or
![\lambda=-1](https://tex.z-dn.net/?f=%5Clambda%3D-1)
. In the first case we arrive at a contradiction (I'll leave establishing that to you). If
![\lambda=-1](https://tex.z-dn.net/?f=%5Clambda%3D-1)
, then we have
![\begin{cases}2x+z=0\\2z+x=0\end{cases}\implies x=0,z=0](https://tex.z-dn.net/?f=%5Cbegin%7Bcases%7D2x%2Bz%3D0%5C%5C2z%2Bx%3D0%5Cend%7Bcases%7D%5Cimplies%20x%3D0%2Cz%3D0)
This means
![y^2=64\implies y=\pm8](https://tex.z-dn.net/?f=y%5E2%3D64%5Cimplies%20y%3D%5Cpm8)
so that the points on the surface closest to the origin are
![(0,\pm8,0)](https://tex.z-dn.net/?f=%280%2C%5Cpm8%2C0%29)
.