How does optimize() work?

[XB G]

Just curious, we define sdpvar variables (e.g. `x`) in the main function, before calling `optimize()`, `value(x)` returns `NaN`. After solving the problem using `optimize()`, `value(x)` gives the optimal value.

How does YALMIP achieve this? 

[XB G]

More specifically, `x` is not input or output of function optimize(). How is `x` (defined in the main function) involved in the operations in `optimize()`?

[Johan Löfberg]

As I've already explained, the name of a variable is completely uninteeresting.

A variable is built from internal variables which have indicies, and these will be associated with values when a problem is solved

yalmip('clear');

banana = sdpvar(1);getvariables(banana)

apple = sdpvar(1);getvariables(apple)

optimize([],(apple-4)^2+(banana-7)^2)

value(recover(1))

value(recover(2))