How does one take care of constraints when setting up an expected returns maximization model of the following kind:
$y=f(x-r)$ if $z > a$,
$z=g(x-s)$ if $w>b$.
I want to max $E=prob\times y + (1-prob)\times z$. Without the constraints, the first order conditions are easy to solve. How do I introduce the inequality constraints into this maximization problem?
Thanks!
