Langrarian multiplier

51 Views Asked by At

Consider the following function
$$f(x, y)=x^4-y^2$$ And Set $A=\{(x,y)\in R^2: x^2+y^2=1\}$ is required.

  1. find the Lagrangian equation that determines the extreme point of $F$ on $A$ and calculates the solution for this equation.
  2. characterize the above equation into local maxima and minima.
  3. using second order condition explain if there are global maxima and minima.
2

There are 2 best solutions below

0
On

$\nabla (x^4 - y^2) = \nabla(\lambda(x^2+y^2 - 1))$

$4x^3 = 2\lambda x\\ -2y = 2\lambda y$

From the second equation we get: $\lambda = -1 \text { or } y = 0$

if $\lambda = -1\\ 4x^3 = -2x\\ x = 0$

Maxima -- $(1,0),(-1,0)$

Minima -- $(0,1),(0,-1)$

0
On

We know that $$\nabla f=(4x^3,-2y)$$and$$c(x,y)=x^2+y^2-1=0$$therefore the Lagrangian would be $$L(x,y)=\nabla f+\lambda\nabla c=(4x^3,-2y)+\lambda(2x,2y)$$also the Hessian of the function is $$H_f=\begin{bmatrix}12x^2&0\\0&-2\end{bmatrix}$$for finding maxima and minima we have $$4x^3=2\lambda x\\-2y=2\lambda y$$if $y=0$ we have $x=\pm 1$ and $\lambda=2$. The only feasible direction in these points is $\vec d=(0,d_2)$ for any $d_2\in\Bbb R-\{0\}$ from the other side$$dH_f(\pm 1,0)d^T=-2d_2^2<0$$which means that the points $(\pm 1,0)$ are local maxima.

If $\lambda=-1$ we obtain $$4x^3=-2x$$ the only possible candidates are $(0,\pm 1)$ that both are local minima similarly. Here is a sketch:

enter image description here