Danny_Kim
(Do-Yup Kim)
1
In many answers, e.g., How to express “x*(2^(1/x)-1)” in CVX

, `exponential(1)`

is used.

What is the difference between `exponential`

and `exp_cone`

?

I have tried to find `exponential`

in CVX manual, but I cannot find it.

There are only `exp_cone`

, defined as

I think to minimize x*(2^{1/x}-1), I have to use the right-most code as follows.

However, in How to express “x*(2^(1/x)-1)” in CVX

, Mark_L_Stone recommended to use the following code: `{1,x,z} == exponential(1)`

Please let me know what the difference is between `exponential`

and `exp_cone`

?

Also, why he recommended to use `==`

instead of `<in>`

?

Thank you.

The CVX Users’ Guide erroneously shows `exp_cone`

. The actual implemented syntax uses `exponential`

, as used in my answers.

I don’t believe

`<in> exp_cone`

works, because CVX does not recognize `<in>`

, at last when I just tried it in CVX 2.2

1 Like

Danny_Kim
(Do-Yup Kim)
3
Here, does `==`

imply “in”? That is, is the phrase `{1,x,z} == exponential(1)`

representing that a tuple (1,x,z) is in a set exp_cone?

The syntax `{1,x,z} == exponential(1)`

works. i don’t know whether any other syntax works.

I didn’t design or develop CVX. I’m just telling you what works. I suggest you use that.

1 Like