Ant Colony Optimization PID Control of Hypnosis With Propofol Using Renyi Permutation Entropy as Controlled Variable

Abstract

General anesthesia is a critical procedure in clinical surgery. To offer a closed control of depth of hypnosis, we establish a closed-loop anesthetic delivery system for propofol anesthesia. The system consists of three components: 1) three-compartment pharmacokinetic model; 2) a pharmacodynamics model obtained by identifying the relationship of effect-site concentration and Renyi permutation entropy via particle swarm optimization; and 3) ant colony optimization proportion integration differentiation controller. The performance of Renyi permutation entropy and bispectral index in tracking the effect-site concentration is evaluated by the prediction probability. The assessments of the performance of the controller are using: 1) the rising time, percent overshoot, and settling time and 2) median performance error, median absolute performance error, wobble, divergence, and integral absolute error. The results show that the prediction probability of Renyi permutation entropy (0.79±0.13 and mean±standard deviation) is higher than the bispectral index (0.74±0.15). The ant colony optimization proportion integration differentiation controller is quick to respond to sudden changes and maintains the stabilty at the desired depth of hypnosis (rise time and overshoot are 4.53±1.96 min and 3.48±1.49 (%), respectively). In conclusion, the proposed closed-loop anesthetic delivery system has potential value for accurate anesthetic administration.

Keywords

• General anesthesia
• closed-loop anesthetic delivery system
• Renyi permutation entropy
• ant colony optimization proportional integral derivative
• particle swarm optimization