Artificial Neural Network(ANN)-based nonlinear optimization of modeling on biomicrofluidic vesicles generation

Biomicrofluidics has been an effective tool to manipulate vesicles at micrometer-scale since the last decade, particularly for the monodisperse microemulsions production. In this work, the flow regime of vesicle generation was studied in the biomicrofluidic environment for the numerical description of microvesicle size variation. Biomi-crofluidic vesicles generation is a complicated process for the study on fluidic silhouette, evolution mechanism, as well as fluidic manipulation at micrometer-scale. Modeling of the descriptive observation permits to understand the inside mechanism of these phenomena. Both linear modeling and optimized nonlinear modeling were introduced. The rough linear model to express vesicle generation was found to be somewhat short of effectiveness. Artificial Neural Network(ANN) technology was applied to perform nonlinear optimization. The results from verification procedures confirmed the improved descriptive quality for this nonlinear model. Besides, to our knowledge, this work for the first time introduced ANN technology to ameliorate vesicle production for pharmaceutical application as well as life science.