Hello all,

I am really confused…some of my friends say that they used cvx for their nonlinear convex programming…

I have a serious question: some days ago I asked for writing x*log(1+x/y) in cvx, and you told me how to write it: rel_entr(x+y,y)+rel_entr(y,x+y) .
But here is my question: CAN we put constraints like x*log(1+x/y) < const. ?

I wrote this constraint on cvx (the whole of problem is linear except this), cvx (under MOSEK academic license) gives this error…:

**Successive approximation method to be employed.**

** For improved efficiency, Mosek is solving the dual problem.**

** Mosek will be called several times to refine the solution.**

** Original size: 860 variables, 300 equality constraints**

** 200 exponentials add 1600 variables, 1000 equality constraints**

**-----------------------------------------------------------------**

** Cones | Errors |**

**Mov/Act | Centering Exp cone Poly cone | Status**

**--------±--------------------------------±--------**

** 0/ 0 | 7.000e+00 0.000e+00 0.000e+00 | Unbounded**

** 0/ 0 | 7.000e+00 0.000e+00 0.000e+00 | Unbounded**

** 0/ 0 | 7.000e+00 0.000e+00 0.000e+00 | Unbounded**

**-----------------------------------------------------------------**

**Status: Infeasible**

**Optimal value (cvx_optval): -Inf**

but without this constraint the problem had been solved correctly.

Thanks alot for your help