The lineartion of bilinear variable.

[MK]

Dear Sir, 

           I have a constraint:

                 xy = 1, (0 <= x <= 1)

           I want to transform it into a linear constraint, and i have tried the McCormick envelope. However, the upper bound of y is infinity, and i am stucked.

         

           Could you give me some suggestions about how to solve it.

           Looking forward to your reply.

Sincerely,

Haibin Lv.

[Johan Löfberg]

First, you canont transform it to a linear constraint. McCormick outer bounds it with a linear set

However, finite bounds are required (and reasonably small for it to make sense)

What kind of application do you have where you cannot bound one of the decision variables?

[MK]

Sorry for delayed reply.

I have mistaken the constraints, and it's get work now.

I am researching on wireless communication optimization.

Thank you for your reply, Johan.