Probabilistic optimal control of blood glucose under uncertainty

MARIANO DE PAULA; ERNESTO C. MARTINEZ

Londres

Congreso; 22nd European Symposium on Computer Aided Process Engineering; 2012

European Federation of Chemical Engineering (EFCE)

Insulin-dependent diabetes mellitus is a chronic disease that requires a careful management of insulin infusion rates. A novel simulation-based approach to probabilistic optimal control of blood glucose using Gaussian Process Dynamic Programming (GPDP) and Bayesian active learning is proposed. GPDP is an approximate value function method that integrates reinforcement learning with Gaussian Processes (GP) for seeking an optimal control policy in the face of an uncertain dynamics. The obtained control policy is compactly represented using Gaussian process hyper-parameters (mean and variance) over a wide range of physiological states to facilitate an “on-a-chip” implementation. Safe adaptation of the simulation-based control policy to a patient-specific lifestyle upon data is proposed. Presented results are very promising.