The Statistical Simplex Method for Experimental Optimization with Process Data

ERNESTO CARLOS MARTINEZ

Barcelona

Congreso; Eurepean Symposium on Computer-Aided Process Engineering, ESCAPE 15; 2005

Federación Europea de Ingeniería Química

Experimental optimization with scarce and noisy process data is a key issue in laboratory automation for faster chemical process research and development, real-time process optimization, extremum-seeking control systems and self-calibrating instruments. To deal successfully with noise and uncontrollable factors in experimental design for process optimization a statistical characterization of a local optimum is proposed. The Kendall?s tau statistic is used for characterizing a local optimum as a cluster center of strongly correlated points. A statistical simplex algorithm that resorts to correlation-based ranking of simplex vertices for reflection, expansion, contraction and shrinking steps is proposed. Results obtained in run-to-run optimization of the operating policy of a semi-batch reactor are presented.