Data-Driven Modeling and Control of Complex Dynamical Systems Arising in Renal Anemia Therapy

TL;DR

This paper presents a data-driven approach combining extended dynamic mode decomposition and model predictive control to efficiently model and control erythropoiesis in anemia therapy, demonstrating comparable results to traditional methods.

Contribution

It introduces a novel application of EDMD and MPC to a complex biological system, enabling efficient control strategies for anemia treatment models.

Findings

EDMD effectively approximates non-linear erythropoiesis dynamics.

MPC based on EDMD achieves control results comparable to traditional methods.

The approach enhances computational efficiency in modeling complex biological systems.

Abstract

This project is based on a mathematical model of erythropoiesis for anemia, which consists of five hyperbolic population equations describing the production of red blood cells under treatment with epoetin-alfa (EPO). Extended dynamic mode decomposition (EDMD) is utilized to approximate the non-linear dynamical systems by linear ones. This allows for efficient and reliable strategies based on a combination of EDMD and model predictive control (MPC), which produces results comparable with the one obtained in past publications for the original model.