| The research presented in this thesis has attempted to go beyond the present state of diabetic modelling by developing a representation that has a strong physiological basis and that consists of key components which have been validated. The model has been developed using data from human diabetic patients subjected to perturbations that resemble those occurring in daily life. We then expect that the structure, even though it contains some empiricisms and rough approximations, will have some applicability to treatment of diabetic individuals.; After close examination of the underlying physiological dynamics and systematic, quantitative analysis, we have been able to reduce the dynamic complexity to a group of smaller, identifiable components which fit together to produce the overall blood glucose response. This modular approach not only makes the modelling more manageable, but also allows for easy implementation of structural changes (e.g. modelling a normal person) or improvements.; The model is in its initial stages of verification and testing. However, it has provided significant information concerning interesting dynamic features which have been observed in real patient situations. |
