Model-Based Insulin Therapy Scheduling: A Mixed-Integer Nonlinear Dynamic Optimization Approach

This article aims at developing a better insulin injection scheduling strategy for diabetes. For this purpose, the subcutaneous (sc) absorption behaviors of available insulin and the overall glucose-insulin dynamics for diabetes are investigated at first. Therein several sets of clinical data from literature are applied to verify the overall glucose-insulin dynamic models through parametric estimation. The problem of searching the optimal injected time, type, and dosage of insulin are then formulated as a mixed-integer nonlinear dynamic program (MINDP). The optimal injection schedules are consequently found for it 24 h cycle in three scenarios by adjusting either the insulin injection times, or insulin types, or insulin dosage, or other combinations of these factors. The corresponding improvement in glycemia control in each scenario is demonstrated. The robustness of suggested therapy schedules to inconsistency of scheduled situations is finally exemplified. It is expected that the proposed optimal therapy scheduling can serve as a valuable reference for physicians and patients to take better glucose control on a daily basis.