Abstract:
A multi objective optimization approach using a Kriging model coupled with a Multi
Objective Genetic Algorithm (MOGA) is applied to a blast damage maximization
problem composed of two objectives, namely number of casualties and damage to
buildings. The predicted Pareto front is located using a MOGA on the Kriging model.
The location with maximum uncertainty along the Pareto front is added to the list of
sample points. After each sampling, the Kriging model is reconstructed and this process
is repeated until the maximum uncertainty is reduced.
The cases run show that the Pareto front is not always intuitively discernable. `Best
locations’ can vary significantly depending on the weight given to each optimization
objective. The results also indicate that the effect of the additional cost incurred by the
procedure to construct the `model of the model’ totally compensates the computational
expense.
