Deep Reinforcement Learning for Closed-Loop Blood Glucose Control: Two Approaches

Reinforcement learning, thanks to the observation-action approach, represents auseful control tool, in particular when the dynamics are characterized by strong non-linearityand complexity. In this sense, it has a natural application in the biological systems field wherethe complexity of the dynamics makes the automatic control particularly challenging. This paperpresents a combined application of neural networks and reinforcement learning, in the so-calledfield of deep reinforcement learning, for the glucose regulation problem in patients with diabetesmellitus. The glucose control problem is solved through the Deep Deterministic Policy Gradient(DDPG) and the Soft Actor-Critic (SAC) algorithms, where the environment exploited for theagent's interactions is represented by a glucose model that is completely unknown to agents.Preliminary results show that the DDPG and SAC agents can suitably control the glucosedynamics, making the proposed approach promising for further investigations. The comparisonbetween the two agents shows a better behaviour of SAC algorithm.