Optimization on Metamodeling-Supported Iterative Decomposition

The recently developed metamodel-based decomposition strategy relies on quantifying the variable correlations of black-box functions so that high dimensional problems are decomposed to smaller sub-problems, before performing optimization. Such a two-step method may miss the global optimum due to its rigidity or requires extra expensive sample points for ensuring adequate decomposition. This work develops a strategy to iteratively decompose high dimensional problems within the optimization process. The sample points used during the optimization are reused to build a metamodel called PCA-HDMR for quantifying the intensities of variable correlations by sensitivity analysis. At every iteration, the predicted intensities of the correlations are updated based on all the evaluated points and a new decomposition scheme is suggested by omitting the weak correlations. Optimization is performed on the iteratively updated sub-problems from decomposition. The proposed strategy is applied for optimization of different benchmark and engineering problems and results are compared to direct optimization of the undecomposed problems using Trust Region Mode Pursuing Sampling method (TRMPS), Genetic Algorithm (GA), and Dividing RECTangles (DIRECT). The results show that except for the category of un-decomposable problems with all or lots of strong (i. e., important) correlations, the proposed strategy effectively improves the accuracy of the optimization results. The advantages of the new strategy in comparison with the previous methods are also discussed.