Genetic fuzzy system modeling and simulation of vascular behaviour

Background: The purpose of our project is to identify the rule sets and their interaction within the framework of cardiovascular function. By an iterative process of computational simulation and experimental work, we strive to mimic the physiological basis for cardiovascular adaptive changes in cardiovascular disease and ultimately improve pharmacotherapy. For this purpose, novel computational approaches incorporating adaptive properties, auto-regulatory control and rule sets will be assessed, properties that are commonly lacking in deterministic models based on differential equations. We hypothesize
that pivotal rule sets governing physiological processes are species independent and could therefore be key to better understanding of translational aspects of cardiovascular function and adaptation in disease. Elucidation of rule sets, their dependencies and interactions may lead to a mechanism-based identification and validation of targets that better translate from a laboratory animal to the human situation and may present a tool for more optimal pharmacotherapy of disease with fewer adverse events. Ultimately this approach could be used not only in cardiovascular function and adaptive behavior but also in principle for any
physiological system that is characterized by auto-regulatory control and adaptation.
Methods: Currently, one modeling approach is being investigated, Genetic Fuzzy System (GFS). In Genetic Fuzzy Systems, the model algorithm mimics the biologic genetic evolutionary process to learn and find the optimal parameters in a Fuzzy Control set that can control the fluctuation of physical features in a blood vessel, based on experimental data (training data). Our solution is to create chromosomes or individuals composed of a sequence of parameters in the fuzzy system and find the best chromosome or individual to define the fuzzy system. The model is implemented by combining the Matlab Genetic algorithm and Fuzzy system toolboxes, respectively. To test the performance of this method, experimental data sets about calculated pressure change in different blood vessels after several chemical treatments are chosen as training and
testing data sets. In the simulation, the fuzzy control system is trained by pressure data of one blood vessel and tested with pressure data of other blood vessels.
Results: Right now, some rough results show that trained fuzzy control system can be used to predict the pressure change of different blood vessels.
Conclusion: Genetic fuzzy system is one of potential modeling methods in modeling and simulation of vascular behavior.