Trying to solve the Hydro-Unit Commitment Problem (HUCP) , the most robust model of the problem, is in the MINLP form.
One (seemingly lazy ) approach is to formulate the problem and let a commercial solvers (say in GAMS) solve it for you.
The bilinear term in HUCP is the power generation function.
Generated power = (efficiency) * ( net head ) * ( turbine flow)
Which is the product of two continuous decision variables.
I wanted to know, if there is a general recommended way to handle this bilinear equality constraint?
If it can be handled efficiently, and the modified problem is now MILP, then the state-of-art solvers ( say MOSEK ) can obtain global optimal solutions.
It is a step further from model robustness to solution robustness maybe.