Respiratory dynamics and hemodynamics modeling by conventional and genetic algorithms

Juan Antonio Woo, University of Texas at El Paso


The focus of this research was to determine the validity of six different respiratory impedance models and a physiologically related rapid-prototyped in-vitro bypass model by using conventional and genetic algorithms. Optimization algorithms, using both normal patient data and diseased patient data from an impulse oscillometry system, were developed in this research to determine the best parameter estimate vectors for each respiratory model. Two models, including a proposed model, were found to perform the best for both sets of data by comparing the estimates obtained from conventional descent-based algorithms against those yielded by the genetic algorithm. The observation that these models perform well for both normal and “abnormal” patient data suggest that more research examining parameter comparisons with respect to different diagnoses is needed. The analysis of a rapid-prototyped bypass model, its transfer function, and a proposed equivalent electric circuit was performed by means of the descent-based and genetic algorithms. The results suggest that a lumped-parameter linearity assumption cannot be used for its characterization. Furthermore, it is concluded that the genetic algorithm is a viable method for physiological model analysis. ^

Subject Area

Engineering, Biomedical|Computer Science

Recommended Citation

Woo, Juan Antonio, "Respiratory dynamics and hemodynamics modeling by conventional and genetic algorithms" (2004). ETD Collection for University of Texas, El Paso. AAIEP10619.