Design of MDIs for type 1 diabetes treatment via rolling horizon cardinality-constrained optimisation

Recent results on cardinality constrained optimisation are used to obtain an optimised Multiple Daily Injection (MDI) insulin regimen for Type 1 Diabetes patients. The optimisation is implemented in a rolling horizon fashion to account for unannounced events. A blood glucose model is described and an algorithm developed to illustrate the aforementioned strategy. The impact of varying the number of allowable injections in a given day is also studied. Real patient data is used to obtain the models and simulation results are included.