```
min 0.5*( x(6)^2.5x(7)^(-0.5) )+0.4( x(8)^2.7x(7)^(-0.6) )
s.t. 0.5(x(1)^(-0.1))(x(2)^(-0.2))(x(3)^(-0.3))(x(4)^(-0.1))(x(5)^(-0.2))-0.6*(x(1)^(0.01))(x(2)^(0.03))(x(3)^(0.02))(x(4)^(0.05))(x(5)^(0.01))<=0.5
0<=x(i)<=2,i=1,2,3,4,5
1<=x(6)<=11
1<=x(7)<=12
1<=x(8)<=13
```

# How to solve the following problem with CVX?

**zhuzhu**(zhuzhu) #1

**Mark_L_Stone**(Mark L. Stone) #2

Have you proven that this is a convex optimization problem. Why isn’t CVX accepting my model? READ THIS FIRST!

I think your objective function is convex. I’ll let you fix up the first constraint with missing symbols, make sure parentheses are correct, etc. Please prove it is convex.

**zhuzhu**(zhuzhu) #3

I’m sorry, there was a sign error in the question, and the following one is correct

min 0.5*( (x(6)^2.5)*(x(7)^(-0.5)) )+0.4(( x(8)^2.7)*(x(7)^(-0.6)) )

s.t. 0.5(x(1)^(-0.1))*(x(2)^(-0.2))*(x(3)^(-0.3))*(x(4)^(-0.1))*(x(5)^(-0.2))-0.6*(x(1)^(0.01))*(x(2)^(0.03))*(x(3)^(0.02))*(x(4)^(0.05))*(x(5)^(0.01))<=0.5

0<=x(i)<=2,i=1,2,3,4,5

1<=x(6)<=11

1<=x(7)<=12

1<=x(8)<=13

**zhuzhu**(zhuzhu) #4

Inspired by theorems 2.1 and 2.2 in a paper, I constructed constraints and objective functions, but I don’t know how to prove that my constraint function is convex function, which is difficult for me。

This paper is http://users.abo.fi/twesterl/selected-publications/9.%20OMS-AL-TW-2009.pdf

**Mark_L_Stone**(Mark L. Stone) #6

Per the link I provided, it’s your obligation to prove convexity. If you have done that, then perhaps readers can offer reformulation tips.

**Mark_L_Stone**(Mark L. Stone) #9

What is `f`

? That doesn’t appear to be left-hand side of the first constraint, whose convexity you need to prove, and which remains very much in doubt, to put it mildly.

**Mark_L_Stone**(Mark L. Stone) #11

Can you please write out clearly what your constraint is? Is that constraint convex?

**zhuzhu**(zhuzhu) #12

My constraint is f(x)+g(x)<=0.5, I have proved that both f(x) and g(x) are convex,so I think the constraint

f(x)+g(x)<=0.5 is also convex.

**Mark_L_Stone**(Mark L. Stone) #13

Formulate it in CVX using gp mode. Form f, and see what type of objects CVX says it is (just enter f on the command line) Similarly with g. if f and g are both generalized posynomials, then `f + g <= constant`

should be accepted by CVX in gp mode. http://cvxr.com/cvx/doc/gp.html#

**zhuzhu**(zhuzhu) #14

Thank you very much. In the constraint f(x)+g(x)<=0.5, g(x) is a negative term, and I need to move g(x) to the right of the inequality if I use gp mode in CVX , so that I get the constraint f(x)<=-g(x)+0.5, however, g(x)+0.5 is not a monomial,can I still use gp mode in such a situation?

**Mark_L_Stone**(Mark L. Stone) #15

I keep losing track of what you are doing, with minus signs coming and going, inside g, in front of g, whatever, so I really don’t know what your constraint is… So I don’t know whether the 2nd term of your left-hand side is going the correct direction. Of course `f(x) + g(x) <= 0.5`

is not the same as `f(x) <= g(x) + 0.5`

. Please refer to my previous post and study the Geometric Programming (gp) chapter linked there.

**zhuzhu**(zhuzhu) #16

I’m sorry, I may not have made my question clear. I mean, both f(x) and g(x) is convex, therofore the constraint f(x)+g(x)<=0.5 is convex, f(x) is a positive monomial, while g(x) is a negative monomial , how to code with cvx to solve the problem?

**Mark_L_Stone**(Mark L. Stone) #17

You’re claiming a negative monomial is convex? That sounds dubious to me. if somehow that is true, i don;t know how to handle it in CVX, and will have to defer to someone else.

**zhuzhu**(zhuzhu) #19

Dear Mark, thank you very much for your help. In the past few days, I have tried other ways to solve this problem, but there is still no progress. This problem is very important to me. Can you help me find a way to solve it? I will wait for your reply with anxiety and expectation.

**Mark_L_Stone**(Mark L. Stone) #20

If you can not produce a formulation acceptable to CVX, you can use another tool whch can handle “general” nonlinear models.