A genetic algorithm tuned optimal controller for glucose regulation in type 1 diabetic subjects

Abstract

An optimal state feedback controller is designed with the objective of minimizing the elevated glucose levels caused by meal intake in Type 1 diabetic subjects, by the minimal infusion of insulin. The states for the controller based on linear quadratic regulator theory are estimated from noisy data using Kalman filter. The controller designed for a physiological relevant mathematical model is coupled with another model for simulating meal dynamics, which converts meal intake into glucose appearance rate in the plasma. The tuning parameters (weighting matrices) of the controller and the design parameters (noise covariance matrices) of the Kalman filter are optimized using genetic algorithm. The controller based on the combined framework of evolutionary computing and state estimated linear quadratic regulator is found to maintain normoglycemia for meal intakes of varying carbohydrate content. The proposed approach addresses noisy output measurement, modeling error and delay in sensor measurement.

A constrained sub‐optimal controller for glucose regulation in type 1 diabetes mellitus

“…An overall performance, however, would depend not only on the type of control technique involved but also on the model used for the patient and the meal. However, with full state feedback, the availability of information about all the states will result in a much improved performance both in terms of deviation from the basal glucose level and the basal infusion rate . Nonetheless, as mentioned earlier, this would come at the cost of much higher complexity arising because of the incorporation of the observer.…”

Marginalization of Irrationality in Decision Making

“…The decision process in Figure 9.2 normally contains a correlation device (Aitken, 1957;Rodgers and Nicewander, 1988;Dowdy and Wearden, 1983), a causal device (Gasking, 1955;Simon and Rescher, 1966;Pearl, 2000;Byrne, 2005) as well as a control device (Abdoos et al, 2011;Achili et al, 2009;Al-Faiz and Sabry, 2012;Farouk, 2012;Ghosh and Gude, 2012). A correlation machine, and was discussed in detail in Chapters 2 and 4, has a function of filling the missing information and many routines such as artificial intelligence have been developed to deal with this problem.…”

“…To illustrate this principle of the marginalization of irrationality, we will invoke the principle of signal to noise ratio that has been widely applied in information theory (Schroeder, 1999;Choma et al, 2003;Russ, 2007;González and Woods, 2008;Raol, 2009). The principle of signal to noise ratio is used to assess the degree through which the signal is corrupted or influenced by noise.…”

Observer‐based control for plasma glucose regulation in type 1 diabetes mellitus patients with unknown input delay

“…10 The advantage of AP over manual insulin introduction is that the patient will have his glucose levels constantly controlled without being severely affected by his activities, such as unplanned meals, exercise, or stress. 11 The development of APs has generated interest in the scientific community, where different models have been formulated to avoid hypoglycemia and hyperglycemia 10 ; personalized model-based algorithms for patients with different insulin responses 12 ; algorithms for detection and classification for different types of diabetes [13][14][15] ; optimal state feedback controllers with a Kalman filter designed to minimize the elevated glucose levels, 16 among others. The Bergman minimal model (BMM) has been popular since it is a simple nonlinear glucose-insulin model that considers physiological parameters such as glucose efficiency, insulin sensitivity, and insulin degradation.…”