Reinforcement learning was shown to be a suitable alternative to classical control algorithms for automatic blood pressure regulation in simulated post cardiac surgery patients.
Does reinforcement learning improve automatic blood pressure regulation compared to classical control in post cardiac surgery patients?
Reinforcement learning models can serve as an alternative to classical PID control for automated blood pressure regulation in post-cardiac surgery patients based on simulated data.
This work treats the problem of blood pressure regulation in post cardiac surgery patients with an improved model and reinforcement learning. The latter is the focus of the present study, specifically showing that it is suitable for the automatic control of blood pressure in post cardiac surgery patients. The context is the direct use of clinical data, with no need for a patient model. Classical control, with PID or Polynomial RST algorithms, has also been used for blood pressure regulation. An existing model of patient is used here and improved. We use a model here for the only purpose of generating pseudo-clinical data for the use of reinforcement learning. The obtained results confirm that this approach can be an alternative to the classical control.
Sandu et al. (Thu,) conducted a other in Post cardiac surgery blood pressure regulation. Reinforcement learning vs. Classical control (PID or Polynomial RST algorithms) was evaluated on Blood pressure regulation. Reinforcement learning was shown to be a suitable alternative to classical control algorithms for automatic blood pressure regulation in simulated post cardiac surgery patients.