$y^2-8 \ln(x+4)\rightarrow$ min,
such that $-x^2 -y^2+9 \geq 0, y \geq 0$
*I have to find all possible optimal points.
*
Lagragian function is:
$L(x,y,γ_1,γ_2) = y^2 - 8\ln(x+4)+γ_1(x^2+y^2-9) + γ_2(-y)$
KKT conditions are the following:
FEASIBILITY:
$-x^2-y^2+9\geq0$
$y\geq0$
STATIONARITY:
$8/(x+4) + 2x γ_1=0$
$2y+2yγ_1-γ_2=0$
NON NEGATIVE MULTIPLIERS:
$γ_1 \geq 0$
$γ_2 \geq 0$
COMPLEMENTARY:
$ γ_1(x^2+y^2-9)=0$
$γ_2(-y) = 0$
So I have to consider 4 cases.
CASE 1)
Both constraints are active:
$$\left\{\begin{array}{l} x^2+y^2-9=0\\ y=0 \end{array}\right.$$ There are two real solutions, but just one point respects non negative multipliers condition: $$\left\{\begin{array}{l} x=-3\\ y=0 \end{array}\right.$$ (-3,0) is a possible optimal point.
I have some difficulties with the other cases.
CASE 2)
The first costraint is non-active, the second constraint is active:
$y=0 , γ_1=0$
So with stationarity conditions I have :
$1/(x+4)=0$
that is not possibile.
CASE 3)
First constraint active, second constrait non-active:
$$\left\{\begin{array}{l}
x^2+y^2-9=0\\
γ_2=0
\end{array}\right.$$
CASE 4)
Both constraint are non-active:
$$\left\{\begin{array}{l}
γ_1=0\\
γ_2=0
\end{array}\right.$$
I dont't understand how to use the KKT conditions to find possible(or not) optimal points in case 2, case 3 and case 4.