Find the point on the sphere x^2 +y^2 + z^2 = 9 that is closest to the point (2,3,4) (presumably using lagrange multipliers, since thats what the chapter entails lol)
You have to minimize the distance between a point on the sphere and the point (2,3,4), with the constraint that the point lies on the sphere.
how do i do that? :(
distance function is: \[\sqrt{(x-a)^{2} +(y-b)^{2} +(z-c)^{2}}\] let a,b,c be point (2,3,4) and x,y,z point on sphere \[z = \sqrt{9-x^{2}-y^{2}}\] \[d = \sqrt{(x-2)^{2}+(y-3)^{2}+(\sqrt{9-x^{2}-y^{2}}-4)^{2}}\] set partial derivatives equal to 0 and solve for x,y
where did you get the last 4 from? and is there a way to also do this with lagrange?
from the point (2,3,4) and probably...the Lagrange multipliers may come into play later in the problem, i've forgotten how to use those...something with matrices
so.. how do i set partials when everything is under a root? :\ and the lagrange multiplier is just: f(x,y,z) - lambda (g(xyz) - k)
you just take the derivative...its a lot of the chain rule :)
would you mind showing me the steps for just one of them, like Fx? the only thing i can think of is \[\sqrt{2(x-2)}\]
do you know the solution? i want to make sure i did it right first...i didn't use langrange multiplier at all so this won't help you for your class too much :(
8/3, according to the book
? it should be a point
just kidding, wrong answer ^.^ the correct one is \[6/\sqrt{29}, 9/\sqrt{29}, 12/\sqrt{29}\]
ok i got the right answer but doing it a different way
yeahhh i get how to use them, just not for this problem :(
hmm try \[f(x,y,z) = \sqrt{(x-2)^{2} +(y-3)^{2}+(z-4)^{2}}\] \[g(x,y,z)-c = x^{2}+y^{2}+z^{2}-9\] then \[\rightarrow \sqrt{(x-2)^{2} +(y-3)^{2}+(z-4)^{2}} +\lambda( x^{2}+y^{2}+z^{2}-9)\]
now you have to take 4 partial derivatives..fun
ohhhhhh okay :o but.. i still don't understand how to even begin going about getting a partial derivative while everything is under the square root :( :(
\[f(x) = \sqrt{u(x)}\] \[f'(x) = \frac{1}{2\sqrt{u(x)}}*u'(x)\] chain rule
i'll do Fx the derivative inside root is just 2(x-2) right multiply that by derivative of sqrt(u) ...where u is that whole big function \[F_x = \frac{2(x-2)}{2\sqrt{u}} +2 \lambda x\]
i'm kind of having a hard time understanding what you're doing, but let me try..is this correct? \[f_y=2(y-3)/2\sqrt{u}\]
+2lambda y
yes
so then \[f_z = 2(z-4)/2\sqrt{u} + 2z \lambda \]
do you know why the sqrt(u) is on bottom? \[d/dx \sqrt{x} = x^{1/2} = \frac{1}{2}x^{-1/2} = \frac{1}{2\sqrt{x}}\]
OHHHHHHHHHH that totally makes so much more sense now LOL
good...also you can simply all of those by cancelling out the 2's
not in lambda part though
wait another question ... so like lets pretend that i did the chain rule but it WASN'T 2(x-2)/2(u) but it was 2(x-2)/(u) right? and then i had to solve for x... i know you're not supposed to touch the inside when doing the chain rule, but when solving for x, after doing the chain rule.. would i be allowed to distribute then?
yes after differentiating you can distribute if you need to
ohh okay! great! thank you sooooo much!!!! i wish i could like.. give you two medals, or fan you again lol
haha no problem...good luck , you may want to review derivative rules before the test
so the denominator is 2sqrt {(x-2)^2+y-3^2+(z-4)^2, right? do i need to write that out for fx fy and fz? and what can i do with it when i'm solving for stuff? :o
its up to you...since its same for each one you could leave it as "u" or some other variable until you have to break it up i think for this problem it just ends up cancelling out so it doesn't matter
wait.. is it supposed to be f(x) - lambda g(x) or f(x) + lambda g(x)?
it says "+" do you use "-" for minimization problems?
hold on let me solve it and i'll find out
i actually have no idea, i just know the formula i've been using is -lambda :o but that's been for finding local max and min, not sure if that counts as minimization =\ and they don't really explain how to do this type of problem in the book -_-"
btw, is this right? \[2x \lambda = (x-2)/\sqrt{u} , x = (x-2)/ 2 \lambda \sqrt{u}\]
ok using "+" works, i get same answers as before looks right...change the sign though
okay! thank you!! :D
your welcome
so... i'm kind of stuck again :x i got x = -x+2 / 2λ√u y = -y+3 / 2λ√u z = - -z+4/2λ√u and am not sure how to proceed =\
sorry :(
haha no problem...well you need to isolate x, right now you have an x on both sides 2 ways to do this: 1. split right side into 2 fractions, then move x term over 2. multiply the denominator back over to left side, then add the x over...factor out the x..
i just tried to do it in my head real quick, and do they both yield the same answer? :o
yes they do...no matter how you do it x can only equal one expression right
haha oh right... -_-" i knew that..
okay so this may be a silly question but ... lets call 2λ√u m? i split it into two to get x = -x/m + 2/m, moved the x over to get x + x/m = 2/m so then i factor it? to get x (1+1/m) = 2/m ? now.. if i divide by 1+1/m .... ... i'm not sure how to do this :X
ahh yes, lot of algebra involved here you need to combine 1+ 1/m into 1 fraction, then to divide, multiply by its reciprocal
then.. would it be... 1m/m + 1/m? o.O
yep
would that make it 2m/m? would the m's cancel out? to just be 2?! :o
uh oh , you are missing something...but yes you will have the m's cancel m/m + 1/m = (m+1)/m reciprocal -> m/(m+1)
OH lol -_-" that was silly haha... so then is x = 2/m+1?
sorry..yes thats correct
yes! okay :D great! you're my hero. x = 2/m+1 , y = 3/m+1, and z = 4/m+1?!
good...now plug them into last equation which is just g(x,y,z)-c and solve for lambda
so i immediately factored out m to get m(-2(2) - 2(3) - 2(4) which is m(-18) = 9 -18 = 9/m, divide both sides by 9 i get -2 = 1/m, or -2 = 1/2λ√u+1 i'm not sure what else i can do to try to get lambda alone :(
what equation are you using? it should be partial wrt lambda --> x^2 +y^2 +z^2 -9 = 0
and be careful...denominator is 1+m so you can't factor out a m
oh nevermind, i see what you did....m is "m+1" :)
wait i'm using g(x)?? not g'x?!? O.O i thought i had to find fx, fy, fz, and then f lamdba =\
oh.. i see what i did wrong -_-"
think , what is f lambda ?
yeah sorry, m = 2λ√u and we'll make n = 2λ√u+1 just so it makes more sense haha.. so plugging everything in i get 4/n + 9/n + 16/n = -9, can i multiply both sides by n?
yes...the -9 should be +9 when you move it over also the denominator is squared...so technically it should be n^2
oh! i totally forgot about the square -_-" so then is it 4 lambda u? which will make it 28 = 36 lambda^2 u?
lets see...you forgot about the +1 ....n = 1+2lambda*sqrt(u) 28 should be 29....4+9+16
i subtracted 1 from both sides, am i not allowed to do that? :(
btw.. does u ever cancel? :o
29 = 9n^2 take square root sqrt29 = 3n sqrt29/3 = n = 1+2lambda*sqrt(u)
then subtract 1
and yes the u will cancel when we plug lambda back in to solve for x,y,z
ohhhhhh okay, order of operations, lol that's right. so then is it: \[-1+\sqrt{29}/6\sqrt{u} = \lambda\]
is that (-1 +sqrt29) ? should be (-3+sqrt29) when you combine fractions...change 1 to 3/3
OH i totally did that at first andthen was like nahh... lol -_-" okay -3+! is the denominator correct? :o
haha yes you got it
YES!!!!! :D :D : D :D omg thank you soooooooooooo much!!!!!!!!! hey, btw, are you a girl or guy? lol
i just realized that sounds creepy lol. i want to write you a testimony and would like to get the sex correct haha
not done yet...still have to plug in lambda to get x,y,z and im a guy...guess i could add that to my profile
haha with a name like dumbcow that's actually pretty hard to guess. it's not fair that plugging lambda in makes things even more complicated
its not too bad...the sqrt(u) will cancel , just be careful with combining fractions and simplifying actually i will say this method turns out to be less messy than my original way
actually, jk, maybe not... is it 1/root(29)? :o
close...haha you have the solutions remember
yeah but i refuse to look at them :o and i'm not plugging into the original equation yet, i'm just trying to solve for x y and z, right?
right x = 2/(1+2lambda*sqrt(u)) now take the expression for lambda and plug it in
so, i must confess, i'm not exactly sure how to properly plug it in haha but this is what i did: 2/2 x (-3+root(29)/6rootu) x root(u)+1 = 2 / -6 +2root(29)6root(u)) x root(u) +1 , i think the 6 and root u cancel out to get 2 / -1+2root(29)+1, the 1's cancel out to get 2/2root(29) and the 2's reduce to get 1/root(29) :o
ok you're mistake is the 6's do not cancel because the whole numerator is (-6+2root29) if it was (-6+6root29) then you could factor out a 6, then 6's cancel in this case, however, the most you can factor out is a 2 --> 2(-3+root29)/6 = (-3+root29)/3
oh...so then it's... but wait.. this is already a denominator.. so what happens with a denominator of a denominator?! i have 2 / (-2+root(29)/3) is that right?
careful with those fractions \[\frac{\sqrt{29}-3}{3} +1 \neq \frac{\sqrt{29}-2}{3}\]
dammit. 2/ (3+root(29)/6) ?!
Note: when you have to divide by a fraction...multiply by its reciprocal
no thats not it...its alright just think get common denominator right change the 1 to 3/3
oh lol wait i had it over 6 for some reason...its 0 ^.^ 2 / (root(29)/6) you said to multiply by the reciprocal? so then should it be 2/6root(29)?
haha its still over 6 :) reciprocal means flip the fraction....root(29)/3 -> 3/root(29)
why.. why is that six.. still there -_-"..... i'm sorry, you must think i'm sooo stupid lol. please don't hate me... wait so .. am i multplying the entire thing by the reciprocal? so like 2/(root29/3) x 3/root29?
or am i mutiplying 2 /root(29)/3 by 3/root(29)/2? lol
Join our real-time social learning platform and learn together with your friends!