My Math Forum Getting all maxima and mimina...

 Calculus Calculus Math Forum

 October 2nd, 2017, 09:27 AM #1 Senior Member   Joined: Jan 2017 From: Toronto Posts: 140 Thanks: 2 Getting all maxima and mimina... The equation 2x^2-4x+y^2 = 16 describes an ellipse in the plane. (a) Find the point(s) on this ellipse closest to the origin. Answer (-2, 0). (b) Find the point(s) on this ellipse farthest from the origin. Answer: (2, 4) and (2, -4). $\displaystyle f(x,y,\lambda) = x^2 + y^2 - \lambda ( 2x^2 - 4x + y^2 - 16 )$ The above equation yields only (2, 4) and (2, -4), but it does not yields (-2, 0) at all.. Would anyone be able to show me how to get (-2, 0) with lagrange multipler?
 October 2nd, 2017, 10:31 AM #2 Senior Member   Joined: May 2016 From: USA Posts: 802 Thanks: 318 Try the correct equation $L(x,\ y,\ \lambda ) = \sqrt{x^2 + y^2} - \lambda (2x^2 - 4x + y^2 - 16).$
October 2nd, 2017, 10:38 AM   #3
Senior Member

Joined: Jan 2017
From: Toronto

Posts: 140
Thanks: 2

It makes no different.
The below yields (2, 4) and (2, -4). It is still unable to get (-2, 0)

Quote:
 Originally Posted by JeffM1 Try the correct equation $L(x,\ y,\ \lambda ) = \sqrt{x^2 + y^2} - \lambda (2x^2 - 4x + y^2 - 16).$

 October 2nd, 2017, 07:09 PM #4 Senior Member   Joined: May 2016 From: USA Posts: 802 Thanks: 318 First off, technically, you cannot represent an ellipse by a single function. That may or may not be relevant, but you cannot be sure whether it is relevant unless you take it into account. Second, the functions representing an ellipse have bounded domains. Again that may or may not be relevant, but you cannot be sure unless you take it into account. Third, you do not need to use Lagrangian multipliers for this problem. The ellipse is described by: $2x^2 - 4x + y^2 = 16.$ So the real functions that are relevant are: $a(x) = y = +\ \sqrt{16 + 4x - 2x^2} \text { such that } -\ 2 \le x \le 4 \text { and}$ $b(x) = y = -\ \sqrt{16 + 4x - 2x^2} \text { such that } -\ 2 \le x \le 4.$ Why those bounds? $\pm \sqrt{16 + 4x - 2x^2} = y \in \mathbb R \implies 16 + 4x - 2x^2 \ge 0 \implies 16 \ge 2x^2 - 4x \implies$ $8 \ge x^2 - 2x \implies 9 \ge x^2 - 2x + 1 \implies 3^2 \ge (x - 1)^2 \implies$ $3 \ge (x - 1) \ge -\ 3 \implies 4 \ge x \ge -\ 2.$ The distance function between the origin and a(x) is: $c(x) = \sqrt{(x - 0)^2 + \left ( +\ \sqrt{16 + 4x - 2x^2} - 0 \right )^2} = \sqrt{ 16 + 4x - x^2 }.$ $c'(x) = \dfrac{4 - 2x}{2\sqrt{ 16 + 4x - x^2 }} \implies c'(x) = 0 \iff x = 2.$ $x = 2 \implies y = \sqrt{16 + 4(2) - 2(2)^2} = \sqrt{16 + 8 - 8} = 4.$ $distance = c(2) = \sqrt{ 16 + 4(2) - (2)^2 } = \sqrt{16 + 8 - 4} = 2\sqrt{5}.$ So we may have a local extremum at (2, 4). The distance function between the origin and b(x) is: $d(x) = \sqrt{(x - 0)^2 + \left ( -\ \sqrt{16 + 4x - 2x^2} - 0 \right )^2} = \sqrt{ 16 + 4x - x^2 }.$ $d'(x) = \dfrac{4 - 2x}{2\sqrt{ 16 + 4x - x^2 }} \implies d'(x) = 0 \iff x = 2.$ $x = 2 \implies y = -\ \sqrt{16 + 4(2) - 2(2)^2} = \sqrt{16 + 8 - 8} = -\ 4.$ $distance = d(2) = \sqrt{ 16 + 4(2) - (2)^2 } = \sqrt{16 + 8 - 4} = 2\sqrt{5}.$ So may have a local extremum at (2, - 4).. So, the distance may be at an extremum at both (2, 4) and (2, - 4). This would be possible only if the locus was a perfect circle centered on the origin, but it is not. What is wrong? WITH A BOUNDED FUNCTION, WE MUST ALSO TESTS THE BOUNDS. So what is c(4)? $c(4) = \sqrt{16 + 4(4) - (4)^2} = \sqrt{16 + 16 - 16} = 4 < 2\sqrt{5}.$ And what is c(- 2)? $c(-\ 2) = \sqrt{16 + 4(-\ 2) - (-\ 2)^2} = \sqrt{16 - 8 - 4} = \sqrt{4} = 2 < 4 < 2\sqrt{5}.$ $b(- 2) = \sqrt{16 + 4(-2) - 2(-\ 2)^2} = \sqrt{16 - 8 - 8} = 0.$ So we have maximal distance at (2, 4) and (2, - 4) and minimal distance at (- 2, 0). Derivatives are not enough with bounded functions. Last edited by JeffM1; October 2nd, 2017 at 07:15 PM.

 Tags maxima, mimina

 Thread Tools Display Modes Linear Mode

 Similar Threads Thread Thread Starter Forum Replies Last Post Volle Calculus 3 July 28th, 2015 01:12 PM Volle Calculus 4 July 21st, 2015 12:48 AM PhizKid Math Software 1 December 26th, 2012 05:11 AM Dontlookback Calculus 2 May 20th, 2010 12:37 AM ArmiAldi Real Analysis 1 March 6th, 2008 04:51 AM

 Contact - Home - Forums - Cryptocurrency Forum - Top