a hybrid model for optimizing the blood glucose control process in children
Subject Areas : Application of artificial intelligence and information technologyfarshad jafarieh 1 , mohammadreza sanaei 2
1 - Department of Information Technology Management, Qa .C., Islamic Azad University, Qazvin, Iran.
2 - Department of Information Technology Management, Qa .C., Islamic Azad University, Qazvin, Iran.
Keywords: Type 1 Diabetes, Particle Swarm Optimization (PSO), Bregman Divergence, Metaheuristic Algorithms, Glucose-Insulin Modeling,
Abstract :
Type 1 diabetes is one of the most common chronic autoimmune diseases in childhood and remains a major challenge due to its complex metabolic nature. Insufficient control of blood glucose can lead to serious short- and long-term complications. Despite continual advancements, multiple daily insulin injections (MDI) are still the primary treatment strategy, yet their performance is limited by the nonlinear dynamics and inherent delays of the glucose-insulin regulatory system. These limitations have increased the demand for intelligent optimization approaches to support more precise and adaptive insulin control.
This study introduces an integrated optimization framework for improving blood glucose regulation in children with type 1 diabetes. The proposed approach is applied to an advanced glucose-insulin model and simultaneously determines the optimal timing and dosage of insulin injections. The hybrid method combines Particle Swarm Optimization (PSO), Extreme Optimization (EO), and Bregman divergence to enhance convergence, accuracy, and stability in the control process. Initially, a dynamic physiological model based on validated mathematical structures is developed. A control scheme is then designed to minimize deviations from the target glucose level. Within this structure, the hybrid metaheuristic algorithm searches for the optimal control parameters, while Bregman divergence improves solution precision and supports theoretical analysis of algorithmic behavior.
Additionally, the research evaluates the performance of other recent metaheuristic algorithms and compares them with the proposed hybrid strategy. All algorithms were implemented in Python and tested using both simulated datasets and real clinical data collected from 49 children with type 1 diabetes. The results confirm that metaheuristic-based controllers significantly reduce glucose fluctuations, accelerate system response, and enhance overall stability compared to conventional control methods.
قوچانی، م.، تهامی، س. 1399. مقایسه شبکه عصبی خودبازگشتی المن و پرسپترون سهلایه در پیشبینی نوسانات قند خون بیماران مبتلا به دیابت نوع یک. فصلنامه علوم دادههای پزشکی، 8(2):صص:56-45.
محمد، ا. 1399. مدل برگمن و کاربرد آن در مدلسازی دینامیک قند خون. نشریه پژوهشهای دیابت ایران، 14(2): صص:57-45.
فرهمند، ح.، دهقانی، ر. 1395. مدلسازی پاسخ گلوکز به انسولین با استفاده از مدل سورنسن. مجله مهندسی پزشکی ایران، 10(1): صص: 31-23.
فرهمند، ح.،. 1398. تحلیل دینامیک انسولین و گلوکز با استفاده از مدل تولیک. نشریه مهندسی پزشکی، 12(3): صص: 76-67.
کالوری، م. 1399. بررسی مدل هوورکا در شبیهسازی عملکرد سیستم بسته انسولین. فصلنامه علوم پزشکی نوین، 6(1): صص: 97-88.
بتمنی، ع.، خداکرمزاده، س. 2019. طراحی مدل UVA-Padova برای تحلیل پایداری سیستم گلوکز خون. مجله فناوریهای نوین در دیابت، 4(2): صص: 110-101.
Ahmad, Z., Leonhardt, S. 2023. Metaheuristics in healthcare. Artificial Intelligence in Medicine, 137, 102456.
Al-Darraji, N., Li, X. 2020. Optimizing diabetes prediction using PSO. Diabetes Research and Clinical Practice, 165, 108231.
American Diabetes Association (ADA). 2023. Standards of medical care in diabetes—2023. Diabetes Care, 46(Suppl. 1), pp. S1–S291.
Ameri, M., Saccomani, M. P., Cobelli, C. 2021. Automated detection of diabetic retinopathy microaneurysms in retinal fluorescein angiography images using local Radon transform. Computers in Biology and Medicine, 135, 104572.
Atkinson, M. A., von Herrath, M., Powers, A. C., Clare-Salzler, M. 2015. Current concepts on the pathogenesis of type 1 diabetes—considerations for attempts to prevent and reverse the disease. Diabetes Care, 38(6), pp. 979–988.
Bahador, A., et al. 2022. Application of data mining techniques for diabetes diagnosis: A predictive model based on multilayer perceptron algorithm. Journal of Biomedical Informatics, 128, 104010.
Bahrami, S., Mehrnia, A., Farahmand, H. 2019. Artificial neural networks for blood glucose prediction in diabetic patients. Computers in Biology and Medicine, 109, pp. 275–285.
Batmeneh, S., Khodakaramzadeh, P. 2019. PID-based control strategies in artificial pancreas systems. Iranian Journal of Control Systems, 15(1), pp. 45–57.
Bellazi, R., Magni, P. 2004. Modeling diabetes with PSO. Artificial Intelligence in Medicine, 32(3), pp. 175–188.
Cahill, L. E., Pan, A., Willett, W. C., Hu, F. B., Rimm, E. B. 2014. Fried-food consumption and risk of type 2 diabetes and coronary artery disease: A prospective study in 2 cohorts of US women and men. The American Journal of Clinical Nutrition, 100(5), pp. 1444–1453.
Cameron, D., Wilson, P., Johnson, M. 2018. Use of insulin pumps in type 1 diabetes treatment: A systematic review. Journal of Medical Technology, 22, pp. 210–225.
Chase, J. G., LeCompte, A. J., Hann, C. E., Lin, J., Pretty, C. G., Shaw, G. M., Wong, X. W. 2018. A model-based glycemic control system for critical care: Development and initial pilot testing. Medical Engineering & Physics, 40, pp. 26–32.
Deb, K., Pratap, A., Agarwal, S., Meyarivan, T. 2002. A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Transactions on Evolutionary Computation, 6(2), pp. 182–197.
Dovelle, F., Russo, G., Rossi, A. 2014. Time-delay control approaches in artificial pancreas systems for type 1 diabetes management. Journal of Biomedical Control, 8(3), pp. 215–227.
Fahlenbach, C., Müller, T., Schmidt, R. 2012. Effect of stress on blood glucose levels in diabetic patients. Diabetes Care, 35(7), pp. 1504–1510.
Farahmand, B., Dehghani, M. 2016. A backstepping approach for blood glucose control of parker system. Paper presented at the 2016 24th Iranian Conference on Electrical Engineering (ICEE), 18 May, pp. 1300–1305.
Farahmand, H., Dehghani, M., Moradi, A. 2017. Fuzzy logic control for insulin delivery in diabetes management: In silico and in vivo studies. Computers in Biology and Medicine, 85, pp. 23–34.
Farahmand, H., Safaei, M., Dehghani, M. 2019. Fuzzy logic control for blood glucose regulation in type 1 diabetes. Artificial Intelligence in Medicine, 99, 101693.
Hernández, G., García, E., González, F. 2018. LMI-based control approaches for blood glucose regulation in diabetes management. Journal of Control Science and Engineering, 2018, Article ID 8343294.
Herroer, M., Tang, F., Li, J., Arvaneh, M. 2022. Enhancing glucose prediction in T1D using wearable sensors and physiological modeling. Sensors, 22(6), 2219.
Hild, K., Fiedler, J., Müller, R. 2022. Linear control methods for glucose regulation in diabetic patients: A review. Biomedical Signal Processing and Control, 72, 103345.
Hildebrandt, P., Chiang, N., Pateekhum, C. 1996. Subcutaneous absorption of insulin. Diabetes/Metabolism Reviews, 12(3), pp. 195–205.
Hood, M., Wilson, R., Corsica, J., Bradley, L., Schwartz, M. 2018. Psychosocial interventions for adults with type 1 diabetes: A systematic review. Diabetes Care, 41(5), pp. 1007–1015.
Hovorka, R., Shojaee-Moradie, F., Carroll, P. V., Chassin, L. J., Gowrie, I. J., Jackson, N. C., ... & Jones, R. H. 2004. Partitioning glucose distribution/transport, disposal, and endogenous production during IVGTT. American Journal of Physiology-Endocrinology and Metabolism, 282(5), pp. E992–E1007.
Ling, C., Rönn, T., Bacos, K. 2012. Epigenetic regulation of insulin resistance in type 2 diabetes. Cell Metabolism, 15(5), pp. 650–660.
Liu, X., Zhang, Y., Wang, J., Li, H. 2022. Optimization of insulin dosing and blood glucose prediction in diabetes management using machine learning techniques. Journal of Diabetes Science and Technology, 16(3), pp. 567–575.
Liu, Y., Zhang, H., Wang, L., Chen, X. 2024. Integration of CGM data into personalized glucose-insulin models for nocturnal hypoglycemia prevention. IEEE Transactions on Biomedical Engineering. Advance online publication.
Lotz, T. F., Chase, J. G., McAuley, K. A., Shaw, G. M., & Hann, C. E. 2010. Monte Carlo analysis of a new model-based method for insulin sensitivity testing. Computer Methods and Programs in Biomedicine, 97(3), pp. 211–219.
Löwen, P., Kircher, M., Wilinska, M. E., Evans, M. L. 2013. Model predictive control for insulin delivery in type 1 diabetes: Recent advances and challenges. Journal of Diabetes Science and Technology, 7(6), pp. 1411–1424.
Lunze, K., Singh, T., Walter, M., Brendel, M. D., & Leonhardt, S. 2013. Blood glucose control algorithms for type 1 diabetic patients: A methodological review. Biomedical Signal Processing and Control. Department of Health and Human Services, Report No.: GGSDRTY1326598.
Makrom, L., et al. 2022. Hybrid neural-fuzzy systems for diabetes diagnosis: Statistical analysis and computational modeling. Computers in Biology and Medicine, 138, 104855.
Riazi, M., Cobelli, C., Wang, H. 2018. Application of fuzzy logic in medical systems: Modeling uncertainty in diabetes management. Journal of Biomedical Engineering, 40(5), pp. 450–462.
Servit, D. 2021. The impact of sex hormones on insulin resistance: A clinical review. Endocrinology and Metabolism, 36(4), pp. 123–135.
Servit, F., Novak, M., Petrik, J. 2021. Genetic and hormonal factors affecting diabetes modeling: The role of estrogen. Endocrinology & Metabolism, 36(4), pp. 385–394.
Sora, N. D., Shashpal, F., Bond, E. A., Jenkins, A. J. 2019. Insulin pumps: Review of technological advancement in diabetes management. The American Journal of the Medical Sciences, 358(5), pp. 326–331.
Sparacino, M. 2020. Challenges and opportunities in continuous glucose monitoring data analysis. Diabetes Technology & Therapeutics, 22(5), pp. 375–386.
