I am interested in finding ways to capture uncertainty in modeling instead of parameters. Figured I would put this thought down. Adding a dummy constant variable to each posynomial may allow us to capture our exposure to individual constraints in a more fundamental way than looking at the sensitivities to the constraint, especially when we see how # of sigma is 'allocated' in a robust solution. May be interesting to see.
I am interested in finding ways to capture uncertainty in modeling instead of parameters. Figured I would put this thought down. Adding a dummy constant variable to each posynomial may allow us to capture our exposure to individual constraints in a more fundamental way than looking at the sensitivities to the constraint, especially when we see how # of sigma is 'allocated' in a robust solution. May be interesting to see.