GAMS World Forum

Hello all,
I am doing multi-objective optimisation in GAMS environment with the use of a weighted sum of objective functions. By solving multi-objective programming, I want to get Pareto-front of my problem. For this purpose, I have to modify the weights of objective functions in a fine step size without any modification in objective functions and constraints. In short, I want to run my optimisation problem cyclically, with the change in weights of objective functions by a fixed step.
So, please tell me how it can be done.
Very-very thanks in advance for helping me.

Hi,

Have a look at meanvar (https://www.gams.com/latest/gamslib_ml/ ... anvar.html). This model (with two objectives: risk and return) creates the Pareto frontier by making one of the objective (risk) a constraint (risk = p) with varying rhs (p) while optimizing the other single objective (return -> max). The model first finds out what the possible range of the risk objective by min/max this single objective. Hope this gives you some inspiration.

-Michael