Evaluating Deep Q-Learning Algorithms for Controlling Blood Glucose in In Silico Type 1 Diabetes
Patients with type 1 diabetes must continually decide how much insulin to inject before each meal to maintain blood glucose levels within a healthy range. Recent research has worked on a solution for this burden, showing the potential of reinforcement learning as an emerging approach for the task of...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-10-01
|
Series: | Diagnostics |
Subjects: | |
Online Access: | https://www.mdpi.com/2075-4418/13/19/3150 |
_version_ | 1797576003063644160 |
---|---|
author | Miguel Tejedor Sigurd Nordtveit Hjerde Jonas Nordhaug Myhre Fred Godtliebsen |
author_facet | Miguel Tejedor Sigurd Nordtveit Hjerde Jonas Nordhaug Myhre Fred Godtliebsen |
author_sort | Miguel Tejedor |
collection | DOAJ |
description | Patients with type 1 diabetes must continually decide how much insulin to inject before each meal to maintain blood glucose levels within a healthy range. Recent research has worked on a solution for this burden, showing the potential of reinforcement learning as an emerging approach for the task of controlling blood glucose levels. In this paper, we test and evaluate several deep Q-learning algorithms for automated and personalized blood glucose regulation in an in silico type 1 diabetes patient with the goal of estimating and delivering proper insulin doses. The proposed algorithms are model-free approaches with no prior information about the patient. We used the Hovorka model with meal variation and carbohydrate counting errors to simulate the patient included in this work. Our experiments compare different deep Q-learning extensions showing promising results controlling blood glucose levels, with some of the proposed algorithms outperforming standard baseline treatment. |
first_indexed | 2024-03-10T21:46:19Z |
format | Article |
id | doaj.art-0d914bd0e69f470d9dc84c56837a2a9c |
institution | Directory Open Access Journal |
issn | 2075-4418 |
language | English |
last_indexed | 2024-03-10T21:46:19Z |
publishDate | 2023-10-01 |
publisher | MDPI AG |
record_format | Article |
series | Diagnostics |
spelling | doaj.art-0d914bd0e69f470d9dc84c56837a2a9c2023-11-19T14:15:30ZengMDPI AGDiagnostics2075-44182023-10-011319315010.3390/diagnostics13193150Evaluating Deep Q-Learning Algorithms for Controlling Blood Glucose in In Silico Type 1 DiabetesMiguel Tejedor0Sigurd Nordtveit Hjerde1Jonas Nordhaug Myhre2Fred Godtliebsen3Norwegian Centre for E-Health Research, P.O. Box 35, N-9038 Tromsø, NorwayFaculty of Science and Technology, Norwegian University of Life Sciences, Postboks 5003 NMBU, 1432 Ås, NorwayNORCE Norwegian Research Centre, Postboks 22, Nygårdstangen, 5838 Bergen, NorwayDepartment of Mathematics and Statistics, UiT—The Arctic University of Norway, P.O. Box 6050 Langnes, N-9037 Tromsø, NorwayPatients with type 1 diabetes must continually decide how much insulin to inject before each meal to maintain blood glucose levels within a healthy range. Recent research has worked on a solution for this burden, showing the potential of reinforcement learning as an emerging approach for the task of controlling blood glucose levels. In this paper, we test and evaluate several deep Q-learning algorithms for automated and personalized blood glucose regulation in an in silico type 1 diabetes patient with the goal of estimating and delivering proper insulin doses. The proposed algorithms are model-free approaches with no prior information about the patient. We used the Hovorka model with meal variation and carbohydrate counting errors to simulate the patient included in this work. Our experiments compare different deep Q-learning extensions showing promising results controlling blood glucose levels, with some of the proposed algorithms outperforming standard baseline treatment.https://www.mdpi.com/2075-4418/13/19/3150reinforcement learningtype 1 diabetesQ-learningdeep learningartificial pancreas |
spellingShingle | Miguel Tejedor Sigurd Nordtveit Hjerde Jonas Nordhaug Myhre Fred Godtliebsen Evaluating Deep Q-Learning Algorithms for Controlling Blood Glucose in In Silico Type 1 Diabetes Diagnostics reinforcement learning type 1 diabetes Q-learning deep learning artificial pancreas |
title | Evaluating Deep Q-Learning Algorithms for Controlling Blood Glucose in In Silico Type 1 Diabetes |
title_full | Evaluating Deep Q-Learning Algorithms for Controlling Blood Glucose in In Silico Type 1 Diabetes |
title_fullStr | Evaluating Deep Q-Learning Algorithms for Controlling Blood Glucose in In Silico Type 1 Diabetes |
title_full_unstemmed | Evaluating Deep Q-Learning Algorithms for Controlling Blood Glucose in In Silico Type 1 Diabetes |
title_short | Evaluating Deep Q-Learning Algorithms for Controlling Blood Glucose in In Silico Type 1 Diabetes |
title_sort | evaluating deep q learning algorithms for controlling blood glucose in in silico type 1 diabetes |
topic | reinforcement learning type 1 diabetes Q-learning deep learning artificial pancreas |
url | https://www.mdpi.com/2075-4418/13/19/3150 |
work_keys_str_mv | AT migueltejedor evaluatingdeepqlearningalgorithmsforcontrollingbloodglucoseininsilicotype1diabetes AT sigurdnordtveithjerde evaluatingdeepqlearningalgorithmsforcontrollingbloodglucoseininsilicotype1diabetes AT jonasnordhaugmyhre evaluatingdeepqlearningalgorithmsforcontrollingbloodglucoseininsilicotype1diabetes AT fredgodtliebsen evaluatingdeepqlearningalgorithmsforcontrollingbloodglucoseininsilicotype1diabetes |