Calc 3: Equation of a line normal to the contour

Join the discussion
Registration is free. Start your own thread to ask a follow-up.
1 reply · 2K views
mtruong1999

Homework Statement


Given g(x,y)=x2 - xy + 2y2, find the equation of the line normal to the contour that passes through the point (1,2).

Homework Equations


Not 100% positive, but the equation to a plane tangent to a function of 3 variables g(x,y,z) is (partial of x)(x - x0) + (partial of y)(y - y0) + (partial of z)(z - z0)= 0
So, I assume that for a function of 2 variables, using this same equation (without the z component, of course) would result in a line tangent to the contour?

The Attempt at a Solution


I am having trouble picturing this. So far I took the partial derivatives with respect to x (which is 2x - y) and with respect to y (which is -x+4y).
What I think I should do is to plug in (1,2) into the partial derivatives, which
would give me 0 for the partial with respect to x, and 7 for the partial with respect to y. Then I plug into the equation 0(x-1)+7(y-2)=0 which I believe is the equation to the line tangent at (1,2)... then I don't know where to go from here...

--or--

Would I just take the partials with respect to x and y at point (1,2) which gives me 0 and 7 respectively then plug them into the symmetric equations for lines? Hmm but the symmetric for the x component would give me and undefined number, so if I wrote it in vector form instead I would get x=1 and y=2+7t which is nicer. Is this the right answer/ did I do this right.
I'm very lost on the appropriate approach to this problem.
 
Physics news on Phys.org
mtruong1999 said:
Then I plug into the equation 0(x-1)+7(y-2)=0 which I believe is the equation to the line tangent at (1,2)... then I don't know where to go from here...
Apart from simplifying it to y = 2, you are done here. However, as you noted before, this is the equation of the tangent plane - not the line normal to the contour. It is the tangent because ##\nabla g## is normal to the contour and the equation ##(\vec x - \vec x_0) \cdot \vec n## describes the plane that passes through ##\vec x_0## and has ##\vec n## as its surface normal.

mtruong1999 said:
Would I just take the partials with respect to x and y at point (1,2) which gives me 0 and 7 respectively then plug them into the symmetric equations for lines? Hmm but the symmetric for the x component would give me and undefined number, so if I wrote it in vector form instead I would get x=1 and y=2+7t which is nicer. Is this the right answer/ did I do this right.
I'm very lost on the appropriate approach to this problem.
On vector form, the equation for a line in the direction ##\vec n## passing through ##\vec x_0## is indeed ##\vec x = \vec x_0 + \vec n t##, where ##t## is a curve parameter. Now, since your normal has an ##x##-component 0, this can just be summarised as the line ##x = 1##, the parametrisation of ##y## really does not matter.