0

Is the following objective infeasible ?

$min: f(a, b, c) = \max(a, b, c) - \min(a, b, c)$

In case of fixing the RHS, function could be minimzed by maximizing the LHS.

I have reformlated the problem,

$max: f(a, b, c) = \min(a, b, c)$

$s.t \ \ \ g(a, b, c) = max(a, b, c), \ $ $is \ minimum$

In many cases, I can't take a step through maximizing $f$ without violating $g$.


I guess there's something wrong with my assumptions, if so what it is?

1 Answers1

0

Clearly, the objective function is nonnegative.

Since you impose the constraint that the $3$ inputs are distinct, they must be positive.

Let $c=0, b = \frac{\epsilon}2, a = \epsilon$ where $\epsilon > 0$, we can see that the problem doesn't have a minimum as $\epsilon$ can be made arbitrarily small.

Siong Thye Goh
  • 149,520
  • 20
  • 88
  • 149