quadratic_over_affine does not accept adding small positivr number to numerator

[JackJack]

Dear Professor Johan 

rhs = y*c1 - ((x0 + y0)/y0)*quadratic_over_affine(y^2, x+y); % no error

rhs = y*c1 - ((x0 + y0)/y0)*quadratic_over_affine((y+0.1)^2, x+y); % cause error Error using quadratic_over_affine
quadratic_over_affine does not support linear of constant terms

Why does this error happens and what are the way to counter it 😢?

Thank you and best regard !

[Johan Löfberg]

lazy coding in an almost never used operator I guess

much better to learn how to derive the socp model from the epigraph model (y+0.1)^2 <= t*(x+y), t>=0, x+y>=0, or just write a simple epigraph model and yalmip will derive the socp, here you could do (y+0.1)^2 <= t*z, t>=0,z>=0, z==x+y and yalmip will detect a simple socp-representable model. the variable z is introduced to make the detection possible

if you absolutely want to use it, you will have to introduce a new variable q to replace y+0.1

[JackJack]

" if you absolutely want to use it, you will have to introduce a new variable q to replace y+0.1 " -> Well this also does not work 

[Johan Löfberg]

I suspect you simply assigned q = y+0.1 instead of defining a new variable q and adding the constraint q == y+0.1

[JackJack]

Thank you that is exactly what happens, after using " adding the constraint q == y+0.1  " it work well, but I have the warning 

Warning: One of the constraints evaluates to a DOUBLE variable 

[Johan Löfberg]

You will have to show the code as that error makes no sense considering what you are saying you do

[JackJack]

Thank you, I know how to do it now