Prof. Dr. H. Hebbel
Empirische Wirtschaftsforschung und Datenverarbeitung
Louvain
Next: Re-simulated
optima for Derringer/Suich
Up: Statistical aspects
of Multi-Criteria-Optimisation
Previous: Sensitivity to
modelling errors:
Design of simulation study
Idea: Use estimated functions as "known world".
- Repeat design of Derringer/Suich several times
- Generate Yi according to estimated
- Estimate new
- Insert estimated
in given desirability functions
- Perform grid search in factor space to find X*opt
If all optima scatter around the "true" optimum: fine!
Dipl.-Stat. Detlef Steuer
1999-10-04