Inverse modeling of biochemical transport

点击翻译按钮获取中文摘要

This proposal addresses current need for computationally efficient inverse models of dispersion processes that are amenable to statistical signal processing. The forward dispersion models have been subject of considerable research interest in the past as they are adequate mathematical representation of diverse processes in various applications: environmental monitoring, homeland security, drug delivery, stock market price fluctuations, atmospheric condition monitoring, etc. Complementary to this approach is inverse modeling in which a set of unknown parameters is estimated using real data measurements. Namely in reality some of the parameters that affect dispersion are unknown (e.g., location of the pollutant leakage, dispersion coefficient for a particular drug, etc.) Due to the complexity of the aforementioned processes and their dependence on a variety of complex parameters and phenomena (e.g. turbulence) inverse models often require significant approximations which in turn affect our ability to accurately estimate unknown parameters and/or predict dispersion process in future.