* 1.731 optimality.gms
* GAMS solution for Kuhn-Tucker example

* suppress unnecessary printout
*$offlisting
$offsymxref
$offsymlist
option
        limrow=0
        limcol=0                ;

variables
        x1               x1
        x2               x2
        y                objective
;

* initialize to insure that GAMS starts with a feasible solution
x1.l = 0.;
x2.l = 1.5  ;

equations
        constraint1             'constraint1'
        constraint2             'constraint2'
        objective               'objective function'
        ;

* use 'power' function for squares since x1 can be negative
constraint1..    x2 =g= 1+power(x1,2)    ;
constraint2..    x2 =l= 2   ;
objective..      y =e= power(x1,2) + power(x2,2)  ;

model optimality /all/                                 ;
solve optimality using nlp minimizing y                ;