Regression of Multicomponent Sticking Probabilities Using a Genetic Algorithm

详细信息    查看全文

• 作者：Ian J. Laurenzi ; John D. Bartels ; Scott L. Diamond
• 刊名：Industrial & Engineering Chemistry Research
• 出版年：2006
• 出版时间：August 2, 2006
• 年：2006
• 卷：45
• 期：16
• 页码：5482 - 5488
• 全文大小：190K
• 年卷期：v.45,no.16(August 2, 2006)
• ISSN：1520-5045

文摘

A genetic algorithm (GA) was developed for the purpose of regressing composition-dependent aggregationkernels from time series of experimentally measured component or size distributions. The GA evolves initiallyrandom populations of kernel models in accordance with the principles of microevolution. To test the robustnessof the GA, functionally diverse kernels-including one describing the shear-mediated aggregation of bloodcells-were constructed. The stochastic time evolution of their corresponding aggregation processes werethen simulated under physiological conditions via Monte Carlo. The GA successfully regressed the kernelsunderlying these "gold standard" datasets-where we employ the term in the sense of a "trusted reference"-from these simulation results, reproducing the multicomponent kernels to a maximum relative deviation ofless than 9% over their entire composition ranges. Finally, ramifications of these cases pertinent to experimentaldesign are considered, including the effects of extreme initial population ratios for multicomponent aggregationexperiments with extreme population ratios.