Digital Twins for Cloud-Integrated Healthcare 5.0

Divya Sasi Latha ^1* , Arya Krishna ² , and Ritika Yamdagni ³

¹ Chulalongkorn Business School, Chulalongkorn University

Bangkok, Thailand ¹

² Keerapat International School, Bang Khen, Bangkok, Thailand

³ NIST International School, Klongtoey-nua Watthana, Bangkok, Thailand

Abstract: This study explores the potential of digital twin technology in the context of healthcare management, with a focus on improving system efficiency, predictive analytics, and operational decision-making. A digital twin framework is formulated and enhanced using a Genetic Algorithm (GA), which is employed to optimize healthcare resource allocation, patient flow, and treatment planning. The framework integrates real-time data from patient monitoring devices, hospital operations, and cloud-based platforms to generate dynamic, virtual replicas of critical healthcare entities. The GA is utilized to iteratively evolve optimal configurations and decision paths within the digital twin model, ensuring adaptive responses to changing clinical and logistical demands. Through simulation and analysis, the study demonstrates the viability of the proposed GA-driven digital twin approach in advancing the goals of healthcare 5.0 enabling a more connected, intelligent, and patient-centric healthcare ecosystem.

Keywords: Digital Twin, Health Informatics, Cloud Computing, Genetic Algorithm

1. Introduction

The global healthcare landscape is undergoing a profound transformation driven by the convergence of emerging technologies and the pressing need for more intelligent, adaptive, and patient-centered care. This transformation is encapsulated in the concept of Healthcare 5.0, which signifies a shift from digitally connected healthcare systems (Healthcare 4.0) to hyper-connected, intelligent, and responsive ecosystems where human expertise collaborates seamlessly with technologies such as Artificial Intelligence (AI), Internet of Things (IoT), 5G, robotics, and cloud computing [1] . At the core of healthcare 5.0 lies the objective of creating a resilient, predictive, and highly personalized healthcare environment that addresses the multifaceted challenges of modern medical practice, including workforce shortages, clinician burnout, and rising operational complexity.

One of the most promising enablers of this next-generation healthcare paradigm is Digital Twin technology. Originally developed for industrial applications, digital twins are now being repurposed for healthcare to create real-time, dynamic virtual representations of physical entities such as patients, hospital equipment, and entire clinical workflows. These digital replicas are continuously updated with live data from IoT sensors, wearable devices, and cloud-based hospital information systems, enabling simulation, monitoring, and optimization of medical operations [2] . By mirroring the physical world in a virtual space, digital twins allow for proactive decision-making, early risk detection, and personalized treatment planning trademarks of the Healthcare 5.0 vision.

However, as healthcare systems increasingly rely on distributed and cloud-based infrastructures, a new challenge emerges ensuring cloud visibility. Cloud visibility refers to the ability to track, monitor, and interpret data flow, resource usage, and operational performance across various digital health platforms. Without adequate visibility, the integration of advanced technologies into clinical workflows can lead to information silos, increased latency, and compromised patient safety [3] . This study addresses this critical gap by proposing a Genetic Algorithm (GA) driven digital twin framework specifically designed to enhance cloud visibility within healthcare 5.0 environments. Genetic Algorithms, inspired by the principles of natural selection and evolution, are powerful optimization tools capable of finding near-optimal solutions in complex, multi-variable systems [4] . Within the context of this research, GAs is used to dynamically evolve configurations within the digital twin model optimizing patient flow, resource allocation, and clinical scheduling while ensuring that data traffic across the cloud is transparent, efficient, and predictable. This integration empowers healthcare providers with the situational awareness needed to deliver timely interventions, reduce system downtime, and improve patient outcomes.

This research is guided by the following key questions:

How can digital twins enhance situational awareness and decision-making in cloud-based healthcare environments?

What role do Genetic Algorithms play in optimizing the performance of digital twin models in dynamic healthcare settings?

What are the measurable impacts on cloud visibility, patient outcomes, and operational efficiency?

The scope of this study focuses on smart hospital systems, digital ICUs, remote diagnostics, and cloud-connected medical devices [5] . The innovation lies in the unique integration of GA based optimization with real-time digital twin simulation, offering a novel and scalable framework for enhancing cloud observability and healthcare intelligence thus contributing significantly to the advancement of healthcare 5.0 research.

The aim of this paper is to develop a comprehensive framework for the integration of digital twins within the domain of healthcare operations, aligned with the principles of Healthcare 5.0. This framework is designed to empower healthcare organizations to systematically leverage digital twin technology for optimizing operational workflows, enhancing clinical decision-making, and improving cloud-based system visibility. By incorporating Genetic Algorithms into the digital twin architecture, the proposed model provides a dynamic and adaptive solution to manage the complexity of modern healthcare environments. The resulting framework offers a scalable, intelligent foundation that supports predictive, personalized, and efficient healthcare delivery overlaying the way for a more connected, resilient, and patient-centric future. The remainder of this paper is organized as follows: Section 2 presents the literature review, outlining relevant research on digital twins and healthcare optimization. Section 3 details the methodology, including the design and implementation of the proposed framework. Section 4 provides a comprehensive discussion of the simulation results and model performance. Section 5 highlights the theoretical and practical implications of the study. Finally, Section 6 concludes the paper with key findings and future research directions.

2. Literature Review

2.1. Evolution of Healthcare Technologies: From Digitization to Intelligent Systems

The evolution of healthcare technologies has historically followed an incremental path, transitioning from manual paper-based systems to digital recordkeeping, and now toward integrated intelligent environments. This trajectory is commonly described through the stages of healthcare 1.0 to healthcare 5.0 [6] . The earliest phase, Healthcare 1.0, relied heavily on manual, reactive care delivery. With the introduction of computers and digital databases, Healthcare 2.0 and 3.0 introduced electronic medical records and basic digital communication systems. Healthcare 4.0 marked a major turning point by incorporating smart devices, IoT, AI, and big data analytics to create more connected and automated healthcare services. This phase enabled real-time data exchange across hospital systems and supported clinical decision-making. However, despite these advancements, Healthcare 4.0 lacked the sophistication required for personalized, predictive, and adaptive care systems. In response, Healthcare 5.0 has emerged as the next paradigm, characterized by seamless human-machine collaboration, decentralized systems, and hyper-connected environments that leverage cloud computing, 5G, robotics, and immersive technologies [7] . The focus has shifted from digitization to intelligent optimization where systems are capable of self-learning, real-time response, and autonomous improvement. Central to this shift is the adoption of digital twin technology, which serves as a virtual counterpart to real-world physical systems and processes in healthcare [8] .

2.2. Digital Twins in Healthcare: Capabilities and Future Potential

The concept of a Digital Twin was first introduced in the manufacturing and aerospace sectors, where digital replicas of physical assets were used for simulation, monitoring, and predictive maintenance. The healthcare industry has since adopted this model, recognizing its transformative potential for personalized medicine, operational efficiency, and dynamic system modelling [9] .

In the healthcare context, digital twins can represent individual patients, hospital operations, or medical devices. These virtual models are continuously updated with real-time data from IoT sensors, wearable technologies, and electronic health records: [10] . They enable physicians to simulate and predict patient responses to treatment, optimize surgical workflows, and manage hospital logistics dynamically. Recent studies highlight the potential of digital twins to support precision healthcare, improve resource allocation, and enable predictive diagnostics. For instance, a digital twin of an ICU can simulate bed occupancy, ventilation settings, and patient deterioration risks, thereby improving response times and treatment efficacy [11] . Moreover, digital twins enhance cloud visibility, a critical requirement in cloud-native healthcare systems. They offer a real-time operational snapshot that supports fault detection, latency analysis, and network performance monitoring, which are essential for system-wide optimization [12] .

2.3. Limitations of Current Digital Twin Frameworks in Healthcare

Despite the acknowledged potential, current implementations of digital twin frameworks in healthcare remain in a nascent stage. Most frameworks focus on isolated use-cases such as predictive modelling for chronic disease management or resource tracking in smart hospitals. These models are often built using static datasets or semi-dynamic simulations that lack real-time responsiveness [13] . One of the major limitations is the lack of integration between digital twins and intelligent optimization algorithms. Existing systems often depend on predefined rules or basic machine learning models that do not scale well in dynamic, multi-objective environments. Furthermore, many frameworks do not account for cloud infrastructure dynamics, making it difficult to monitor system load, bandwidth efficiency, or data transmission failures. Another challenge lies in the fragmentation of healthcare data. Patient data, sensor readings, and administrative workflows are often siloed across different platforms. While some digital twin systems attempt to unify these data sources, most fall short in providing interoperable, cloud-synced, and security compliant integration. Additionally, scalability and generalizability are key concerns. Digital twin implementations are often tailored to specific hospital environments or medical conditions, limiting their applicability across diverse healthcare settings. This restricts real-world deployment and hinders broader impact [14] . Finally, privacy and data governance remain under-addressed in most existing frameworks. Given the sensitive nature of healthcare data, digital twin systems must include built-in cybersecurity measures and robust access control mechanisms to comply with regulations such as HIPAA and GDPR. However, most current models lack embedded privacy-preserving mechanisms [15] .

2.4. Bridging the Gaps with Genetic Algorithms: Research Contribution

Nearly address the limitations of existing digital twin implementations in healthcare, this study proposes a novel and unified framework that integrates Genetic Algorithms with real-time digital twin systems, aimed at optimizing healthcare operations and enhancing cloud visibility. GAs are a class of adaptive heuristic search algorithms inspired by the principles of natural evolution selection, crossover, and mutation [16] . These algorithms are particularly well-suited for solving complex, multi-objective optimization problems, which are common in dynamic healthcare environments. Traditional digital twin systems, though powerful in simulation and monitoring, often rely on static rules or conventional machine learning models that are insufficient for dealing with the evolving and unpredictable nature of real-world healthcare scenarios. In contrast, the integration of GAs introduces a dynamic optimization capability that enables the digital twin not only to reflect ongoing processes but also to iteratively improve them over time based on feedback and system performance [17] . Within this proposed framework, GA is deployed to continuously evolve optimal strategies for Resource Allocation : Efficiently assigning limited healthcare resources such as ICU beds, surgical rooms, and staff to meet patient demands. Patient Flow Optimization: Streamlining patient movement across departments to reduce congestion, waiting times, and care delays. Personalized Treatment Planning: Tailoring intervention plans using real-time patient data from the physical healthcare twin, thus enabling precision medicine. Cloud Resource Balancing: Dynamically distributing cloud workloads to minimize data latency, avoid network congestion, and ensure real-time responsiveness. The Genetic Algorithm component simulates thousands of potential system configurations and selects the most efficient ones based on pre-defined fitness criteria, such as throughput, system load, patient outcomes, and failure tolerance [18] . This makes the system self-improving, resilient, and context-aware qualities that are essential in the fast-paced, high-stakes domain of healthcare. Despite the emergence of digital twin models in various sectors, a critical gap persists in the literature and practice: few studies have successfully developed real-time, optimization-driven digital twin frameworks that provide comprehensive visibility into cloud-integrated healthcare systems. Existing models often remain conceptual, lack cloud observability, or are designed for narrow applications without scalability across diverse healthcare settings. This research addresses the above limitations by presenting a unified, adaptive digital twin framework that comprises three core pillars: Real-Time Digital Twin Simulation: A continuously updated virtual replica of healthcare processes, patients, and infrastructure, dynamically synchronized with real-world data. Cloud-Integrated Monitoring Dashboards: Interfaces that enable transparent monitoring of key system metrics, including data latency, fault detection, resource utilization, and system responsiveness. Genetic Algorithm-Based Optimization Modules: Embedded AI mechanisms that provide adaptive and predictive capabilities for proactive healthcare delivery. By combining simulation, intelligent optimization, and cloud observability into one cohesive architecture, this research contributes a novel, scalable solution to advance the objectives of healthcare 5.0. It supports the vision of a hyper-connected, patient-centric, and operationally efficient healthcare ecosystem, where technology and clinical expertise work in harmony to deliver adaptive, real-time, and sustainable care solutions.

In brief, the literature underscores the transformative potential of digital twin technology in advancing Healthcare 5.0, particularly in enabling real-time decision-making, personalised care, and system-wide optimisation. However, significant gaps remain in existing frameworks, especially regarding real-time responsiveness, intelligent optimisation, and cloud infrastructure visibility [19] . The integration of Genetic Algorithms within digital twin architectures presents a promising solution to address these limitations by providing adaptive, data-driven strategies for healthcare management. Building upon these insights, the following methodology section presents the design and implementation of a framework that incorporates digital twin simulation, and GA-based optimisation aimed at achieving a resilient, scalable, and intelligent healthcare system.

Figure 1 illustrates the proposed framework integrates a physical healthcare twin and a digital twin system within a closed-loop architecture for healthcare 5.0. It visually represents the data-driven loop between real-world healthcare systems and their digital counterparts for real-time performance monitoring, adaptive decision-making, and intelligent system evolution [20] . The physical healthcare twin captures real-world data from patients, devices, and workflows through data collection and analysis using IoT, RFID, and smart sensors. The digital twin system mirrors the physical environment, powered by a GA for dynamic optimization of resource allocation, staff scheduling, and treatment planning, alongside a real-time monitoring module. The process involves iterative testing, improvement, and mapping back to the physical system, ensuring real-time updates and intelligent system evolution. This feedback-driven model enables real-time synchronization, AI-driven adaptability, cloud visibility, and patient-centric optimization. By bridging the gap between physical and digital healthcare environments, the framework supports resilient, transparent, and intelligent decision-making central to the goals of healthcare 5.0.

Figure 1: Digital Twin Approach

3. Methodology

The methodology of this study is designed to fulfil the central objective of developing a Genetic Algorithm (GA)-enabled Digital Twin (DT) framework for optimizing healthcare operations and enhancing cloud visibility in a Healthcare 5.0 environment. The first methodological objective focuses on the design and implementation of a comprehensive digital twin framework that replicates real-time healthcare entities, including patients, hospital devices, and clinical workflows. This virtual environment is dynamically linked with real-time cloud-synced data from IoT sensors, wearable health devices, and electronic health records. It is structured into three modular layers: the patient twin, representing physiological and clinical parameters of individuals; the operational twin, which models resource usage and hospital logistics; and the process twin, simulating care workflows such as diagnostics, triage, and surgical scheduling. This framework aims to provide healthcare organizations with a holistic and interactive model that enables accurate simulation and responsive decision-making in alignment with healthcare 5.0 principles [21] . The second objective is to integrate Genetic Algorithms into the digital twin environment to serve as a powerful optimization engine. These algorithms mimic natural selection processes to solve complex, multi-variable problems inherent in healthcare systems, such as resource allocation, treatment planning, and service prioritization. Within the digital twin, the GA operates by evolving a population of potential solutions based on performance metrics referred to as fitness functions that reflect healthcare goals like reduced patient waiting time, improved care throughput, balanced resource distribution, and minimized system downtime. Key GA operations, including selection, crossover, and mutation, allow the system to iteratively learn and adapt to evolving hospital conditions [22] . This integration ensures that the digital twin is not merely a passive replica but an intelligent, adaptive tool that optimises healthcare delivery in real time. A third critical objective of the methodology is to establish and monitor cloud visibility metrics that offer insights into the performance and reliability of cloud-based healthcare systems. Cloud visibility, a cornerstone of healthcare 5.0 infrastructure, enables transparent system monitoring, fault detection, and performance analysis. The research focuses on four primary metrics: data latency, which evaluates delays in information transfer between sensors, devices, and cloud platforms; anomaly detection rate, which measures the system’s ability to identify unusual patterns or failures; fault recovery time, which captures the duration required to identify and resolve system-level disruptions; and throughput efficiency, which assesses the volume and speed of data processed within the cloud system. These metrics are essential for ensuring that the cloud infrastructure supporting the digital twin remains reliable, scalable, and efficient, particularly in high-stakes healthcare environments. The final methodological objective is to simulate and evaluate the proposed GA-integrated digital twin framework through scenario-based testing and data-driven validation. This includes constructing realistic simulation environments that mirror conditions such as emergency room congestion, ICU overload, or hospital-wide workflow disruptions. Synthetic patient datasets and hospital process data will be used to create comparative performance scenarios between standard (non-optimised) systems and the GA-enhanced digital twin. Evaluation will be conducted using a set of operational and clinical performance indicators, including average treatment duration, patient throughput, system response time, and overall resource utilization. Quantitative validation techniques such as t-tests and ANOVA will be employed to statistically assess the improvements enabled by the framework. Furthermore, sensitivity analyses will be conducted to evaluate the robustness of the Genetic Algorithm under varying conditions, ensuring the model's adaptability and generalizability.

In conclude, the methodological framework of this study is anchored in four core objectives: building a real-time, cloud-integrated digital twin system; embedding a Genetic Algorithm for intelligent optimization; defining and measuring cloud visibility to ensure system transparency; and conducting rigorous simulations to validate operational and clinical improvements. These objectives are interdependent and collectively aim to provide a scalable, intelligent, and data-driven healthcare model aligned with the transformative goals of healthcare 5.0. By systematically addressing both technological and operational dimensions, the proposed methodology establishes a strong foundation for advancing next-generation healthcare delivery systems.

3.1 Case Study Context and Fundamental Assumptions

To demonstrate the feasibility and practical application of the proposed GA enabled Digital Twin framework, this study employs a case study-based methodology situated within a simulated smart hospital environment. The focal point of this case study is a digital intensive care unit (ICU), a critical subsystem of hospital infrastructure where clinical complexity, resource limitations, and the need for timely decisions make it an ideal candidate for optimization through digital technologies. The digital ICU represents a high-density, high-dependency environment that necessitates constant monitoring, adaptive scheduling, and efficient patient flow management, all of which are pivotal to improving overall system performance. The rationale behind selecting the ICU lies in the fact that it embodies the broader challenges faced by smart hospitals: fluctuating patient severity levels, dynamic treatment protocols, constrained human and material resources, and the need for real-time operational decisions. Furthermore, the ICU serves as a data-rich environment, making it suitable for constructing a Digital Twin model that can mirror real-world clinical operations in real time [23] .

The development of the digital twin framework is predicated on several fundamental assumptions that are necessary to simulate a real-world smart hospital scenario under controlled and measurable conditions. First, it is assumed that real-time data streams from the physical ICU are readily accessible through a robust network of IoT enabled monitoring devices. These devices include vital sign monitors (e.g., ECG, oxygen saturation, blood pressure sensors), environmental sensors (e.g., ambient temperature and humidity), and operational sensors (e.g., bed occupancy and ventilator usage). The integration of these data sources forms the basis for creating an accurate and responsive virtual representation of the physical environment. Second, the model assumes that the hospital infrastructure is cloud enabled, facilitating bidirectional communication between the physical healthcare system and its digital counterpart. This cloud infrastructure is expected to support seamless data flow, synchronization, and feedback mechanisms that allow the digital twin not only to monitor but also to inform decision-making in real-time. The cloud platform also supports data storage, analytics, and the deployment of optimization algorithms, including the Genetic Algorithm used in this framework [24] .

The study operates under the assumption that the operational objectives of the ICU are clearly defined and quantifiable. These objectives include metrics such as minimizing patient wait times, maximizing bed turnover rates, reducing staff idle time, and ensuring optimal utilization of critical care equipment. These goals serve as the foundation for constructing objective functions within the optimization model. Defining these parameters upfront ensures that the Genetic Algorithm can operate within a structured fitness landscape to identify the most effective operational configurations. Finally, it is assumed that the digital infrastructure adheres to interoperability standards, allowing data exchange across devices and systems from different manufacturers and departments. This is critical for ensuring the accuracy and completeness of the digital twin simulation, particularly in multi-device environments common to modern ICUs. In summary, the case study is designed to reflect the operational realities of a digital ICU within a smart hospital and is grounded in assumptions that enable the development of a fully functional, cloud-integrated digital twin. These foundational elements guide the subsequent phases of system modelling, algorithm design, and performance evaluation, which are detailed in the sections that follow. First, real-time data streams from IoT-enabled smart sensors embedded throughout the ICU such as vital signs monitors, bed occupancy sensors, and device trackers are assumed to be continuously available. This data stream forms the dynamic backbone of the digital twin, ensuring it accurately mirrors and responds to patient and operational conditions in real time. Second, it is assumed that the hospital's cloud infrastructure supports bidirectional communication, enabling not only real-time data transmission from the physical ICU to the digital twin but also feedback from the digital twin back to the physical environment in the form of optimization insights and actionable recommendations. This two-way interaction is crucial for maintaining a closed feedback loop and real-time adaptability. Finally, the framework assumes that operational goals such as reducing patient wait times, optimizing ICU throughput, and improving equipment utilization are well-defined and quantifiable. These objectives are formulated as fitness functions within the Genetic Algorithm, allowing for precise, data-driven optimization of ICU workflows and resource management. Together, these assumptions establish the technological and operational foundation for building a responsive and intelligent Healthcare 5.0 system.

3.2. Data Collection

To enable the design and simulation of the proposed Digital Twin framework, this study leverages a combination of open-access healthcare datasets and real-time IoT simulation platforms. These platforms serve as reliable sources for both historical and synthetic real-time data that are essential for modelling ICU operations, patient behavior, and infrastructure dynamics in a Healthcare 5.0 environment. The collected data is continuously fed into the cloud backend of the digital twin, enabling real-time synchronization, simulation accuracy, and optimization through Genetic Algorithms. For patient-level clinical data, this study utilizes the MIMIC-IV (Medical Information Mart for Intensive Care, version 4) [25] dataset, maintained by the Massachusetts Institute of Technology’s Laboratory for Computational Physiology. The dataset provides detailed variables such as admission/discharge timestamps, vital signs (e.g., blood pressure, heart rate, oxygen saturation), medication usage, and diagnostic codes, all of which are critical for patient flow modelling and clinical decision simulation in the digital twin environment.

To model healthcare workflows and hospital operations, this study incorporates structured data from platforms such as Simulacrum (UK Health Data Research), HospitalRun (open-source hospital management system data), and HealthData.gov (U.S. Department of Health and Human Services). These platforms offer datasets related to medical staff scheduling, equipment utilisation rates, department-level workflow analytics, and healthcare delivery timeframes, allowing the construction of process-oriented models in the digital twin. In terms of real-time device telemetry and environmental conditions, this study integrates simulated IoT data using platforms such as ThingSpeak (MathWorks IoT analytics platform), OpenIoT (EU-funded open-source middleware for IoT cloud integration), and IoTStream.io. These platforms provide support for live-streaming of virtual sensor data, including continuous monitoring of heart rate, respiratory rate, temperature, SpO₂, and device operation cycles. Through these systems, real-time physiological trends and equipment status are mimicked to enable dynamic patient monitoring and situational awareness within the digital twin. The integration of these datasets focuses on key input variables such as: Patient flow metrics (e.g., admission/discharge timestamps, treatment duration). This hybrid approach to data collection ensures that the virtual model remains operationally aligned with the physical ICU context, allowing it to function as a continuous, adaptive decision-support system within the Healthcare 5.0 paradigm [26] .

Min F(x) =α1​Wt​ + α2​It ​+ α3​Dt​ + α4​Cp​ + α5​Er​ (1)

where: W _t = total patient wait time, It​ = system idle time , Dt​ = total treatment deviation, Cp​ = penalty for unserved critical patients, Er​ = penalty for insufficient emergency bed reservation , α1+α2+α3+α4+α5=1

Table 1 . The derived parameters for ICU digital twin optimization Model

Table 2. The constraints

Equation (1) represents the ICU Digital Twin Optimization Model aims to minimize the overall operational inefficiency by optimizing a multi-objective cost function ​, where the components represent wait time, idle time, treatment deviation, criticality penalties, and emergency reservation penalties respectively. Each parameter is derived from real-time healthcare data, including patient conditions, resource availability, and system dynamics. For instance, Wt captures cumulative patient delays, Dt reflects deviations from optimal treatment durations based on severity, while Cp and Er​ impose penalties for unmet clinical priorities and insufficient emergency capacity. The model operates within defined clinical and operational constraints to ensure realism and safety. These include bed capacity limits, staff-to-patient ratios, treatment time regulations, equipment availability, ICU skill mix, staff working hours, emergency bed buffers, and room type appropriateness. Collectively, the parameters and constraints guide the Genetic Algorithm to generate optimized schedules, resource allocations, and care pathways aligned with healthcare 5.0 principles of intelligence, adaptability, and patient-centric care.

3.3 Model Testing

The ICU Digital Twin Optimization Model is illustrated in Figure 2 , which presents a 30-day simulation output generated using a MATLAB-based environment [27] . Each group of bars represents the day-wise variation of five critical performance indicators: Patient Wait Time (Wt), System Idle Time (It), Treatment Deviation (Dt), Criticality Penalty (Cp), and Emergency Reservation Penalty (Er). These parameters are key components of the multi-objective cost function F(x), continuously computed and visualized through the digital twin system. Following its development, the model underwent rigorous testing and validation to verify its fidelity in replicating real-world healthcare operations. The digital twin was evaluated for its ability to accurately simulate ICU workflows, resource allocation patterns, and patient treatment pathways under dynamic and constrained conditions. This simulation-driven testing phase confirmed that the model could reliably reflect the intricate interdependencies of operational factors within a smart healthcare environment. By analyzing trends and fluctuations in the plotted metrics, stakeholders can identify bottlenecks, predict system overloads, and proactively adjust strategies. Visualization serves as a powerful decision-support tool, ensuring that the optimization model aligns with clinical goals and healthcare 5.0 objectives, such as real-time responsiveness, patient-centric care, and cloud-integrated visibility.

Figure 2: ICU digital twin optimization model for 30 days

3.4 Disruption Management and Genetic algorithm Performance

The GA performance for ICU Digital Twin Optimization is outlines structured , iterative process for improving operational efficiency in smart healthcare environments. The process begins with the generation of an initial population of candidate solutions, each representing a unique configuration of ICU parameters such as bed assignment, staff scheduling, and treatment durations. These solutions are evaluated using a fitness function derived from the digital twin's objective model, which incorporates multi-dimensional performance metrics including patient wait time, system idle time, treatment deviation, criticality penalties, and emergency bed allocation penalties. Through the selection phase, the most promising solutions are chosen based on fitness scores to participate in crossover, where pairs of solutions exchange components to produce potentially superior offspring [28] . A mutation step then introduces variability to the offspring, enhancing the algorithm’s ability to explore diverse solution spaces and avoid local optima. The cycle continues through a convergence check, assessing whether the optimization goals have been met, or the maximum number of iterations has been reached. Upon convergence, the algorithm outputs the optimal solution, which can be directly applied to the ICU digital twin to simulate and execute real-time operational improvements.

This GA-based framework ensures adaptability, resilience, and performance alignment in dynamic healthcare scenarios, fulfilling core principles of the healthcare 5.0 vision. Key Parameters for our digital twin disruption management using GA configuration adopted in this study has been carefully designed to balance exploration and convergence within the ICU Digital Twin Optimization Model. A population size of 500 ensures sufficient diversity among candidate solutions, enhancing the algorithm’s ability to search a wide solution space effectively. A relatively high mutation rate of 0.5 is employed to introduce variability and avoid premature convergence, especially crucial in dynamic and constraint-driven healthcare environments. The crossover probability is set at 0.9, facilitating robust recombination of high-performing traits from parent solutions, which accelerates convergence towards optimal configurations. To preserve the best solutions across generations, an elitism strategy is applied, retaining the top 1% or 5 elite individuals unchanged in each cycle. This helps safeguard against regression in solution quality. The algorithm runs for a maximum of 100 generations, providing adequate iterations to fine-tune solutions while maintaining computational efficiency. Together, these parameters form a strong evolutionary foundation for enabling the digital twin to adaptively optimize ICU resource allocation, staff scheduling, and patient care delivery in real time [29] .

4.Discussions

The bar chart figure 3 illustrates the performance of the ICU Digital Twin Optimization Model under both baseline and disrupted conditions across a 30-day simulation period, with a specific focus on three key penalty metrics: treatment deviation (Dt​), criticality penalty (Cp​), and emergency reservation penalty (Er​). A vertically axis line on Day the onset of a simulated disruption, representing a scenario such as a sudden surge in critically ill patients or an abrupt resource shortfall. Post-disruption, there is a noticeable increase in treatment deviation (dark blue bars) compared to the baseline (light blue), indicating delays or mismatches in treatment durations, potentially due to overburdened staff or unavailable equipment. Similarly, the criticality penalty (red bars) spikes significantly on Days 16 to 19, reflecting that high-severity patients were not treated within clinically acceptable timelines. The emergency reservation penalty (yellow bars) emerges only after Day 15, confirming a failure to maintain the required buffer of ICU beds for emergencies during the disrupted state. These shifts demonstrate the tangible impact of disruption on operational stability. However, the gradual stabilization of these penalties by Day 25 highlights the role of the Genetic Algorithm in adapting the system through dynamic reallocation and optimization. Overall, this analysis underscores the capability of a GA-enhanced digital twin to detect, respond to, and recover from healthcare system disruptions in real time making it a critical component of resilient, intelligent care delivery in the Healthcare 5.0 paradigm.

Figure 3. Impact of Disruption on ICU Digital Twin Optimization Metrics Over 30 Days

5. Theoretical and practical implications

The theoretical and practical implications of this study are both substantial and forward-looking, particularly in the context of evolving healthcare 5.0 paradigms. Theoretically, the research advances the foundational understanding of digital twin technology by framing it not merely as a monitoring tool but as an intelligent, algorithm-driven decision-support system. The formulation of a digital twin framework that integrates real-time data with predictive and optimization capabilities illustrates how such systems can model complex healthcare dynamics. This contributes to academic discourse on cyber-physical systems, healthcare informatics, and smart hospital design by demonstrating how virtual replicas of healthcare processes can be systematically optimized through algorithmic intelligence, such as Genetic Algorithms. Practically, the study offers a tangible blueprint for how digital twins can be deployed to enhance healthcare operations in real-world settings. The framework provides actionable insights for hospital administrators, clinical teams, and system integrators by demonstrating improvements in resource allocation, staff scheduling, patient flow, and emergency readiness. Through dynamic simulation and real-time optimization, the model supports proactive care strategies and operational agility, especially during high-stress scenarios such as ICU overload or systemic disruptions. Furthermore, the integration with cloud infrastructure and data analytics platforms underscores its scalability and compatibility with existing digital health ecosystems. Overall, the implications reinforce the role of digital twins as essential enablers of hyper-connected, patient-centric, and sustainable healthcare systems aligned with the transformative vision of healthcare 5.0.

6. Conclusion

This study has demonstrated the strategic value of digital twin technology as a transformative enabler in modern healthcare management. By formulating and implementing a comprehensive digital twin framework, the research addresses core challenges of operational complexity, resource inefficiency, and lack of system-wide visibility within healthcare ecosystems. The proposed model integrated with real-time patient monitoring data, cloud-based analytics, and intelligent optimization through Genetic Algorithms enables simulation, prediction, and dynamic reconfiguration of clinical workflows and resource allocation. Through detailed simulations, the study confirms that digital twins significantly enhance operational efficiency, support data-driven clinical decision-making, and improve system resilience under both normal and disrupted conditions. This capacity to visualize and optimize care pathways in real time establishes digital twins as essential tools for proactive, rather than reactive, healthcare management. Aligned with the vision of healthcare 5.0, the study reinforces the relevance of intelligent, hyper-connected technologies in creating patient-centric, adaptive, and sustainable healthcare systems. The framework’s ability to support predictive analytics, automate routine decisions, and enable continuous system learning positions it as a foundational element for next-generation smart hospitals. This work not only validates the theoretical promise of digital twin systems in healthcare but also provides practical pathways for their implementation concrete the way for smarter, more resilient, and more responsive healthcare delivery in the digital age.

Reference

[1] H. R. Chi, M. de Fátima Domingues, H. Zhu, C. Li, K. Kojima, and A. Radwan, “Healthcare 5.0: In the perspective of consumer Internet-of-Things-based fog/cloud computing,” IEEE Trans. Consum. Electron., vol. 69, no. 4, pp. 745–755, 2023.

[2] H. Yadav, P. Yadav, N. Joshi, and S. C. Sharma, “Smart healthcare: Paradigm shift in Industry 5.0 using AI,” in Recent Trends in Artificial Intelligence Towards a Smart World: Applications in Industries and Sectors, Singapore: Springer Nature Singapore, 2024, pp. 67–97.

[3] G. L. Santos, P. T. Endo, M. F. F. da S. L. Tigre, L. G. F. da Silva, D. Sadok, J. Kelner, and T. Lynn, “Analyzing the availability and performance of an e-health system integrated with edge, fog and cloud infrastructures,” J. Cloud Comput., vol. 7, pp. 1–22, 2018.

[4] Z. Amiri, A. Heidari, M. Zavvar, N. J. Navimipour, and M. Esmaeilpour, “The applications of nature-inspired algorithms in Internet of Things-based healthcare service: A systematic literature review,” Trans. Emerg. Telecommun. Technol., vol. 35, no. 6, e4969, 2024.

[5] W. Trevena et al., “Model-driven engineering for digital twins: A graph model-based patient simulation application,” Front. Physiol., vol. 15, 1424931, 2024.

[6] X. Zhong et al., “A multidisciplinary approach to the development of digital twin models of critical care delivery in intensive care units,” Int. J. Prod. Res., vol. 60, no. 13, pp. 4197–4213, 2022.

[7] G. Gopal, C. Suter-Crazzolara, L. Toldo, and W. Eberhardt, “Digital transformation in healthcare architectures of present and future information technologies,” Clin. Chem. Lab. Med., vol. 57, no. 3, pp. 328–335, 2019.

[8] K. A. B. Ahmad et al., “Emerging trends and evolutions for smart city healthcare systems,” Sustain. Cities Soc., vol. 80, 103695, 2022.

[9] K. Katsaliaki and S. Kumar, “The past, present, and future of the healthcare delivery system through digitalization,” IEEE Eng. Manag. Rev., vol. 50, no. 4, pp. 21–33, 2022.

[10] O. S. Kobyakova, V. I. Starodubov, N. G. Kurakova, and L. A. Tsvetkova, “Digital twins in healthcare: an assessment of technological and practical prospects,” Ann. Russ. Acad. Med. Sci., vol. 76, no. 5, pp. 476–487, 2021.

[11] A. Haleem, M. Javaid, R. P. Singh, and R. Suman, “Exploring the revolution in healthcare systems through the applications of digital twin technology,” Biomed. Technol., vol. 4, pp. 28–38, 2023.

[12] S. Mihai et al., “Digital twins: A survey on enabling technologies, challenges, trends and future prospects,” IEEE Commun. Surveys Tuts., vol. 24, no. 4, pp. 2255–2291, 2022.

[13] K. Bruynseels, F. Santoni de Sio, and J. Van den Hoven, “Digital twins in health care: ethical implications of an emerging engineering paradigm,” Front. Genet., vol. 9, 31, 2018.

[14] T. Sun, X. He, and Z. Li, “Digital twin in healthcare: Recent updates and challenges,” Digital Health, vol. 9, 20552076221149651, 2023.

[15] R. Sahal, S. H. Alsamhi, and K. N. Brown, “Personal digital twin: a close look into the present and a step towards the future of personalised healthcare industry,” Sensors, vol. 22, no. 15, 5918, 2022.

[16] H. Elayan, M. Aloqaily, and M. Guizani, “Digital twin for intelligent context-aware IoT healthcare systems,” IEEE Internet Things J., vol. 8, no. 23, pp. 16749–16757, 2021.

[17] V. Srinivas and K. Ramanjaneyulu, “An integrated approach for optimum design of bridge decks using genetic algorithms and artificial neural networks,” Adv. Eng. Softw., vol. 38, no. 7, pp. 475–487, 2007.

[18] B. A. Shapiro, Y. G. Yingling, W. Kasprzak, and E. Bindewald, “Bridging the gap in RNA structure prediction,” Curr. Opin. Struct. Biol., vol. 17, no. 2, pp. 157–165, 2007.

[19] Z. Vasicek, “Bridging the gap between evolvable hardware and industry using cartesian genetic programming,” in Inspired by Nature: Essays Presented to Julian F. Miller on the Occasion of his 60th Birthday, Cham, Switzerland: Springer, 2017, pp. 39–55.

[20] L. Wright and S. Davidson, “How to tell the difference between a model and a digital twin,” Adv. Model. Simul. Eng. Sci., vol. 7, pp. 1–13, 2020.

[21] M. Dietz and G. Pernul, “Digital twin: Empowering enterprises towards a system-of-systems approach,” Bus. Inf. Syst. Eng., vol. 62, pp. 179–184, 2020.

[22] B. Sharma, D. Kaushal, M. Sharma, S. Joshi, S. Gopal, and P. Gupta, “Integration of AI, digital twin and Internet of Medical Things (IoMT) for Healthcare 5.0: A bibliometric analysis,” in Proc. 2023 Int. Conf. Adv. Comput., Commun. Inf. Technol. (ICAICCIT), 2023, pp. 1313–1318.

[23] E. M. Eumi, “A systematic review of digital twins in efficient pandemic management with challenges and emerging trends,” Decision Analytics J., 100502, 2024.

[24] Y. Tai et al., “Digital-twin-enabled IoMT system for surgical simulation using rAC-GAN,” IEEE Internet Things J., vol. 9, no. 21, pp. 20918–20931, 2022.

[25] S. Adibi, A. Rajabifard, D. Shojaei, and N. Wickramasinghe, “Enhancing healthcare through sensor-enabled digital twins in smart environments: A comprehensive analysis,” Sensors, vol. 24, no. 9, 2793, 2024.

[26] A. Johnson et al., “MIMIC-IV,” PhysioNet, 2020. [Online]. Available: https://physionet.org/content/mimiciv/1.0 / (accessed Aug. 23, 2021).

[27] H. Tamaki, H. Kita, and S. Kobayashi, “Multi-objective optimization by genetic algorithms: A review,” in Proc. IEEE Int. Conf. Evol. Comput., 1996, pp. 517–522.

[28] D. Bansal, Real-Time Data Acquisition in Human Physiology: Real-Time Acquisition, Processing, and Interpretation A MATLAB-Based Approach. Cambridge, MA, USA: Academic Press, 2021.

[29] J. C. Cebrian and N. Kagan, “Reconfiguration of distribution networks to minimize loss and disruption costs using genetic algorithms,” Electr. Power Syst. Res., vol. 80, no. 1, pp. 53–62, 2010.