New approach to inequalities!

Algebra Level 4

Find the global minimum of the function f ( x , y ) = x 4 + y 4 50 x 2 968 y 2 + 234882 f(x,y)=x^4+y^4-50x^2-968y^2+234882 .

If you come to the conclusion no such minimum exists enter 666.


The answer is 1.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

2 solutions

We can rewrite the given function as ( x 2 25 ) 2 + ( y 2 484 ) 2 + 1 { ({ { x }^{ 2 }-25 }) }^{ 2 }+{ \left( { y }^{ 2 }-484 \right) }^{ 2 }+1 . Thus, the global minimum is 1 1 , when x = 5 , 5 x=5, -5 and y = 22 , 22 y=22, -22 .

Indeed the intended solution, there is a lengthier one with calculus but completing he square is awesome.

Aditya Narayan Sharma - 5 years, 1 month ago

A really good approach

Shreyansh Mukhopadhyay - 3 years, 4 months ago
Pi Han Goh
Apr 29, 2016

Relevant wiki: Second Derivative Test

Let f x f_x and f y f_y denote the partial derivatives of f ( x , y ) f(x,y) , respectively, with respect to x x and y y .

Then at extrema, f x = 4 x 3 100 x = 0 f_x = 4x^3 - 100x = 0 , f y = 4 y 3 1936 y = 0 f_y = 4y^3 - 1936 y = 0 . Solving them gives x = 0 , ± 5 x = 0, \pm 5 , and y = 0 , ± 22 y = 0, \pm 22 .

Similarly, let f x x f_{xx} and f y y f_{yy} denote the second partial derivatives of f ( x , y ) f(x,y) , respectively, with respect to x x and y y .

Then, f x x = 12 x 2 100 f_{xx} = 12x^2 - 100 and f y y = 12 y 2 1936 f_{yy} = 12y^2 - 1936 .

We apply the theorem for the second derivative test discriminant as D : = f x x f y y ( f x y ) 2 D := f_{xx} f_{yy} - (f_{xy})^2 .

Let ( x 0 , y 0 ) (x_0, y_0) denore the point of the local extremum.

If D > 0 D> 0 , then f x x ( x 0 , y 0 ) f_{xx} (x_0, y_0) is a local minimum.
If D < 0 D< 0 , then f x x ( x 0 , y 0 ) f_{xx} (x_0, y_0) is a local maximum.
If D = 0 D=0 , the point is a saddle point.

Then D = 1 2 2 ( x 2 100 12 ) ( y 2 1936 12 ) > 0 D = 12^2 \left( x^2 - \dfrac{100}{12} \right) \left(y^2 - \dfrac{1936}{12} \right) > 0 is a local minimum when ( x 2 , y 2 ) = ( 25 , 484 ) (x^2,y^2 ) = ( 25, 484 ) .

Thus f ( x , y ) f(x,y) has a minimum value at ( x 2 , y 2 ) = ( 25 , 484 ) (x^2,y^2) = (25, 484 ) , and the value of this number is 2 5 2 + 48 4 2 50 ( 25 ) 968 ( 484 ) + 234882 = 1 25^2 + 484^2 - 50(25)-968(484)+234882 =\boxed1 .

Yeah I was waiting for a calculus solution. Does this verify it has global minimum rather than local ?

Aditya Narayan Sharma - 5 years, 1 month ago

Log in to reply

Oh curses! I read your question as "local minimum" instead of "global minimum".

Summoning Comrade @Otto Bretscher for guidance.

Pi Han Goh - 5 years, 1 month ago

Log in to reply

Yes Comrade's solution is awaited. By the way in case of the global minimum I have a favourite way. Use the Hessian . A few days ago I had a brief dialouge with comrade @Otto Bretscher . There was linear terms in his question where we were asked to find the same, so I have avoided linear terms here. Infact both global minima and maxima exist for the function

Aditya Narayan Sharma - 5 years, 1 month ago

First let me mention that you do not quite get the second derivative test right (the role of D D ). See here .

You don't need multivariable calculus here since the variables are "uncoupled." You can separately use single variable calculus to find the global minima of x 4 50 x 2 x^4-50x^2 (attained at x = ± 5 x=\pm5 ) and of y 4 968 y 2 y^4-968y^2 (attained at y = ± 22 y=\pm 22 ) and then infer that you get the global minimum of f ( x , y ) f(x,y) at ( x , y ) = ( ± 5 , ± 22 ) (x,y)=(\pm 5,\pm 22) .

I would say that in this case completing the squares is clearly the most efficient approach.

Otto Bretscher - 5 years, 1 month ago

Log in to reply

Here 'uncoupled' referrs that if we don't have terms involving both x,y that is extrema of one is independent of other we don't really need to perform derivative tests ? Oh so it's just a matter of finding minima of two different expressions. Thanks Comrade !

Aditya Narayan Sharma - 5 years, 1 month ago

Log in to reply

@Aditya Narayan Sharma Yes, exactly. We do perform (single-variable) derivative tests, but one variable at a time.

When we deal with diagonal matrices (the Jacobian matrix and the Hessian matrix), we might as well just work with the diagonal entries... it saves some ink ;)

Otto Bretscher - 5 years, 1 month ago

Log in to reply

@Otto Bretscher Yes definitely , well I am trying to grasp the concept of those circulant matrices and even matrices of all kind. They seem to caome handy in lot of cases

Aditya Narayan Sharma - 5 years, 1 month ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...