The optimization problem formulation , as given by (2), also
had to find metamodel parameters,\(\beta_{0},\beta_{1}\ \text{and}\ \beta_{2}\). An additional
constraint to this linear programming formulation included a minimum
and maximum value for each of the parameter estimates to vary in of
[-1000, 1000]. The parameters for the Solver included:
- The use of multiple starting points using a population of size 100.
- A convergence of 1x10-4.
- Bounds required on variables.
- A level of precision of \(1\times 10^{-3}\) for the functions
Sphere, Rosenbrock, Griewank, Goldstein-Price, Easom, and Schwefel; a
level of precision of \(1\times 10^{-9}\) for the function
Rastrigin.