0
$\begingroup$

In applying Kuhn-Tucker theorem, I encountered a problem of the form $\max_{0 \leq x \leq a}f(x)$. If there is no upper bound for $x$, then the Kuhn-Tucker result gives the first order conditions for the maximization: if $x$ is a maximizer, then $f'(x) \leq 0$ and $xf'(x) = 0$, with the constraint $x \geq 0$, of course. However, I did not see how to work the bound $a$ in here. Need help.

| cite | improve this question | | | | |
$\endgroup$
0
$\begingroup$

The KKT conditions require

$$f^\prime (x) = \mu-\lambda$$ $$ \lambda x = 0 $$ $$ \mu(x-a) = 0 $$

where $\lambda$ and $\mu$ are non-negative KKT multipliers for the constraints $x\ge 0$ and $x \le a$ respectively.

If there were no upper bound $a$ (or, equivalently, if the constraint $x\le a$ did not bind), then the relevant constraints described above reduce to the first-order conditions you state.

| cite | improve this answer | | | | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.