The normal line through a point P (not through the origin) on the parabola y = x^2 will intersect the parabola at another point Q. Find the coordinates of the point P for which the distance from P to Q is minimized.
and what are your thoughts?
I now that I need to have the points P and Q on x^2. I also know that I need an equation and then need to differentiate and set the differentiated equation to zero so that I can get the minimized value. I don't know how to obtain the initial equation though.
we need a line .... and a line requires a slope and a point. we have a point, how can we determine the slope at that point?
we take the derivative at that point
correct, and the slope that is perpendicular to a tangent, is called normal .... given a tangent of y'(a), the normal slope is -1/y'(a) does that make sense? what is our point when x=a? what is our normal slope when x=a?
yes that makes sense. So the point at x = a would be (a,a^2) and the normal slope would be -1/(2a), right?
now we have a line y= -1/(2a) (x-a) +a^2 we have the parabola:y=x^2 they meet at x=a, and x= the other point. how do we find that other point?
we set that equation equal to x^2 ?
yep x^2 = -1/(2a)(x-a)+a^2 x^2 + 1/(2a)(x-a) = a^2 x^2 + 1/(2a)(x-a) - a^2= 0 this is just a quadratic that can be solved ... so what do we come up for the other point?
wait, how do you solve for x in that equation? There's an x in the denominator, and I don't really know how the quadratic formula would work in that case.
the x is not in the denominator, but we can expand it out thru distribution and rewrok it into something like this x^2 + x/(2a) -(a^2+1/2) = 0 2a x^2 + x -(2a^3 + a) = 0 \[x=\frac{-1\pm\sqrt{1+4(2a^3+a)}}{4a}\]
so that would be the x value of point Q?
i wonder if we can simplify that any ... or if ive made an error in the process. as long as we can define the 2 points, we can find the distance between them
\[y=-\frac{1}{2a}(x-a)+a^2\] \[y=-\frac{x}{2a}+a^2+\frac12\] \[x^2+\frac{x}{2a}=a^2+\frac12\] x=a is obviously one solution
\[2ax^2+x-a(2a^2+1)=0\] \[x=\frac{-1\pm\sqrt{1+4a(2a^2+1)}}{4a}\] well,we have our 2 point regardless \[(x_1,x^2_1)~(x_2,x^2_2)\] both in terms of a \[d^2=(x_1-x_2)^2+(x^2_1-x^2_2 )^2\]
finding the derivatives with respect to a might be fun :)
the basic premise then takes us to : \[d^2 = n^2 + m^2\] \[2dd'=2nn'+2mm'\] \[d'=\frac{2nn'+2mm'}{2d}=0\] \[nn'+mm'=0\]
is d' dd/da, so you're taking the derivative with respect to a?
with respect to a, yes seeing how our distances rely upon x, when x=a ... the whole system depends on a i beleive
ok, so would the result that you got with implicit differentiation: nn' + mm' = 0 still work when you don't really have to differentiate implicitly in the regular problem?
yes, but the nm stuff was just generalizing it
because everything in the regular problem is in terms of a, so would that equate to: \[n \frac{ dn }{ da } + m \frac{ dm }{ da }\] ?
yes
spose n = x1 - x2 n = 1/2a^(-1) +4a^2 +2 n' = -1/2a^(-2) +8a
keep in mind that this is just a thought that is in my head for a solution, and there may exist simpler approaches :)
ok, and whatever value I get after I solve the equation should be the minimum distance between P and Q. right?
yes :) spose we let the points be defined as (a,a^2) and (f(a),[f(a)]^2) d^2 = (f(a)-a)^2 + ([f(a)]^2 - a^2)^2 2dd' = 2(f(a)-a)(f'(a)-a') + 2([f(a)]^2 - a^2)(2f(a) f'(a) - 2a) dd' = (f(a)-a) (f'(a)-a') + ([f(a)]^2 - a^2) (2f(a) f'(a) - 2a) d is irrelevant, or it may help to simplify when we divide it ... dunno, but when d'=0 we get 0 = (f(a)-a) (f'(a)-1) + ([f(a)]^2 - a^2) (2f(a) f'(a) - 2a) maybe we can solve this for a?and thereby determine the points p and q?
just another thought, may or not be worthwhile
ok, thank you so much more your time and your help! I really appreciate it!
youre welcome. good luck, and see what you can do to improve the method if you can.
Thanks!
Join our real-time social learning platform and learn together with your friends!