Robust modelling of biological neuroregulators

Abstract

The modelling of biological neuroregulated systems is usually complex because of their non-structured and distributed architecture. Software agents are a very useful tool to use in the modelling of this kind of systems, allowing increasing the model with characteristics which are not present in the biological equivalent. This work proposes a model of a neural regulator incorporating modularity, flexibility and scalability. The main approach has been the modelling of each neural centre as an agent, incorporating the distributed behaviour of the system. On the other hand, the model can solve local dysfunctions in the centres using extra diagnostic information. The neural regulator of the lower urinary tract has been implemented as an example. We have developed several experiments adding artificial dysfunctions to the model and comparing the results with a normal functioning of the model.