Originally posted by: SystemAdmin

Before getting to use CPLEX for solving optimization problems, I thought it could handle all convex optimization problems, not just restricted to quadratic/linear ones. In my opt. problem, the objective function is like c1*x + c2*x^2 + c3*exp(-x), where c1, c2 and c3 are nonnegative and x>=0. Obviously, the obj. function is convex, but CPLEX cannot handle it, right? I mean there's an exponential term.

This is more like a mathematical problem, but any suggestions would be cool. dunno but CPLEX might surprise me!

Thanks!
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: gangulo

The exponential function is almost conic-quadratic representable, thus you may be able to get an accurate approximation with CPLEX by including additional variables and second-order cone constraints. Check this presentation from Arkadi Nemirovski's page, in particular slides 24-25.

I hope it helps!
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: WongXR

I am running into the same problems,would you tell me more detail please?THX

#CPLEXOptimizers
#DecisionOptimization