Fuzzy percolation model for loss of consciousness under acceleration stress

In science and engineering, models of real systems are constructed to identify underlying mechanisms and properties, to make predictions or to control their behavior. In the majority of cases, the system to be modeled is based on abstract concepts, on parameters difficult or impossible to precisely quantize, and on experiments limited by technology or the system itself. Due to the unavoidable presence of uncertainty and the usual need to simplify the system and reduce its computational complexity prior to modeling, engineers need to make reasonable assumptions based on their knowledge of the system dynamics. The fuzzy percolation model presented here is an extension to the percolation model (crisp model) introduced in [3]. Both models were built using the tools and concepts of complexity theory by mapping the reticular activating system (RAS) in the brain into a percolation network whose nodes represent a non-specific area of the central nervous system (CNS). To construct the percolation model, data from neuroscience, medical science and aviation medicine were used to determine the local node parameters. Both models utilize the same structure, parameters and variables to represent the cardiovascular and nervous systems; the crisp percolation model makes reasonable assumptions about not well known parameters without taking into account the inherent uncertainty and unavoidable imprecision in their representation. In this article, we show how this uncertainty can be modeled using fuzzy systems. We then test the model and compare its performance to the crisp model using human data. Since both, crisp and fuzzy models were able to duplicate human studies with a high degree of fidelity, we conclude that a precise description of the parameters is not necessary when modeling a complex system, provided the parameters remain within physiologically reasonable bounds .