Hernández-Rivera A. , VELARDE RUEDA, PABLO ANIBAL, Zafra-Cabeza A. , Maestre J.M.
No
Rev. Iberoam. Autom. Inform. Ind.
Article
Científica
01/01/2025
001525039900010
2-s2.0-105002411521
Mathematical models of biomedical systems can help practitioners design safer and more effective drug administration cycles. To achieve this goal, the mathematical model of tumoral growth and the impact of chemotherapy are used in the decision-making process. However, biomedical systems are prone to a high degree of uncertainty, not only from measurement errors but also from unmodeled dynamics of the system and interpatient variability. To address this issue, probabilistic constraints have been applied to the control of the drug administration process, making it more robust against disturbances. This work compares a non-linear and a linearized version of the stochastic formulations of the model predictive control. Both algorithms enhance treatment efficacy and safety, with differences in conservativeness and computational cost. © 2025 Universidad Politecnica de Valencia. All rights reserved.
Controlled drug delivery; Model predictive control; Optimal control systems; Pharmacokinetics; Predictive control systems; Stochastic control systems; Stochastic models; Stochastic systems; Targeted drug delivery; Biomedical systems; Decision-making process; Drug administration; Nonlinear model predictive control; Nonlinear predictive control; Pharmacokinetic and drug delivery; Predictive control; Predictive control strategy; Stochastic optimal control; Stochastics; Chemotherapy