I got the equation for three responses.
Now various parameters are affecting the various responses in different manner.
If i will use GRA then the % contibution is not matching
I want Global unique solution which optimize all three responses, in context of three variables as input .
I think you should use Genetic Algorithm or ANN for this purpose..
I would suggest to read this question:
https://www.researchgate.net/post/what_is_the_best_way_for_solving_non-linear_optimization_problems_when_you_are_looking_for_global_optimum
It has really nice and detailed explanations in it.
I think you need weighted non-linear regression. That’s mean you must minimize
SumiSumj(yijexp-yijmod)2wij where yijexp are the experimental value of variable j at experiment i and wij are the weights for balance the % of contribution. You will not find a unique solution but the most appropriated value for wij are: wij=(1/varij) and you must estimate the variance of each response (varij). If you have replicates of experimental points you can use they, if you have not replicates you can estimate the varij from the parameters variance in first estimation of parameters.
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology