To explore the perspectives of healthcare professionals and the experience of families concerning post-PICU follow-up to identify key points for the further development of these programs. Qualitative study. PICUs in France involved in post-PICU patient follow-up. Sixteen healthcare professionals involved in post-PICU follow-up at 11 centers, and 18 family participants in enrolled in longitudinal post-PICU follow-up at a tertiary center, took part in semi-structured interviews and observations from January 2022 to June 2024. None. Three main themes emerged. First, therapeutic alliance as a cornerstone: families stressed the importance of maintaining relationships with PICU professionals throughout follow-up, which facilitated recovery and helped them deal with periods of doubt. Second, professional commitment despite institutional constraints: healthcare providers were highly motivated to maintain follow-up programs despite limited resources. Third, barriers to follow-up engagement: families reported a financial burden, and emotional challenges related to traumatic memories during consultations. Our study highlights the need for structured support for post-PICU follow-up programs that take account of both institutional constraints and family needs. The work suggests that maintenance of the therapeutic relationships established during the PICU stay is crucial for successful follow-up. Future guidelines should address resource allocation and accessibility while preserving the human dimension of care.

Abstract: Health economic evaluations address resource allocation problems and have been widely used in health technology assessment for adopting new interventions. Numerous countries and international societies have developed economic evaluation guidelines. There is consensus among the guidelines on several key principles including the type of analysis [cost-utility analysis (CUA) preferred], time horizon of the analysis (long enough required), health outcome [the quality-adjusted life-year (QALY) preferred], and sensitivity analysis (various analyses recommended). Long-term CUAs which aim to capture all intended effects and unintended side effects of interventions, ideally over the lifetime horizon, are generally recommended. However, strict application of these recommendations could introduce greater uncertainty in economic evaluation results. We examine various guidelines to distill the commonalities and suggest more nuanced approaches for improving the presentation of the credibility of economic model results (i.e., the likelihood of cost-effectiveness results being trusted). We propose several considerations for enhancing the credibility of economic evaluation conclusions. First, when it is challenging to make a reliable projection of long-term outcomes, health economists may limit the time horizon to a shorter time supported by the available data. Second, the economic evaluation can be conducted using natural units of health when high-quality utility data are not available. Third, health economists need to understand whether the available data can support complex disease models so they can choose a proper level of model complexity. Fourth, health economists need to specify factors that are associated with health benefits, and to control remaining factors (e.g., background mortality and baseline characteristics). Fifth, instead of focusing on selecting fitted parameters from real-world data, economic evaluations should use the best quality clinical evidence for their key model inputs. In conclusion, economic evaluations should aim to produce the most credible cost-effectiveness estimates by adhering to rigorous methodological standards and by adapting recommendations to suit the specific context. Health economists should clearly identify and report potential biases, assumptions, the quality of evidence informing key model parameters, and sources of uncertainty to ensure the credibility of the cost-effectiveness results and their use in decision and policy making.

As specialty palliative care (SPC) programs expand nationwide, variability in referral criteria and care scope has led to inconsistent care delivery and confusion among clinicians and patients. As demand for SPC increases, workforce limitations necessitate prioritization frameworks. The Brookdale Department of Geriatrics and Palliative Medicine at the Mount Sinai Health System convened a task force of six palliative care physicians to reach consensus on SPC scope of practice. The group conducted a literature review, surveyed department clinicians, and solicited input from 10 peer academic institutions to inform guideline development. The task force met four times over three months and finalized the guideline through departmental review and leadership endorsement. Findings highlighted wide variation in definitions of "serious illness" and appropriateness for SPC. Most surveyed clinicians supported a definition requiring both high mortality risk and negative impact on quality of life or caregiver burden. The resulting institutional guideline emphasizes prioritization of patients with serious illness and high risk of mortality, explicitly excluding patients with chronic pain or psychosocial distress in the absence of a serious illness. The guideline also addresses safe opioid prescribing and recommends tracking "non-eligible" referrals to identify unmet system needs. This initiative offers the first SPC scope guideline. Ongoing discussions, both at individual institutions and nationally, may be necessary to determine the importance of consistency in defining and communicating the scope of SPC. Health care leaders can use this guideline to address resource allocation, health policy, workforce education, and public understanding of palliative care.

Rheumatic heart disease (RHD) is the most prevalent acquired heart condition among individuals under the age of 25, characterized by a significant risk of cardiac morbidity and mortality, which contributes to premature death. This study utilizes data from the Global Burden of Disease (GBD) 2021 to present an updated assessment of the burden of RHD at national and sub-national levels within Iran. This study is a systematic analysis using data covering the years 1990 to 2021. The age-standardized incidence rate (ASIR), age-standardized prevalence rate (ASPR), age-standardized death rate (ASDR), and age-standardized disability-adjusted life years (DALYs) per 100,000 population for RHD in Iran and its 31 provinces were extracted from the GBD database for both sexes across age groups ranging from under 5 years to over 95 years. A temporal trend analysis and subnational assessment were conducted using DisMod-MR version 2.1. The ASPR and ASIR per 100,000 population remained relatively stable from 1990 to 2021. In contrast, the ASDR per 100,000 population experienced a significant decline, decreasing from 3.3 (95% uncertainty interval [UI]: 2.4 to 4.5) in 1990 to 1.4 (1.1 to 1.6) in 2021, corresponding to a reduction of 58.8% (-73.0 to -39.8). The ASDR per 100,000 population for women decreased by 57.5% (-78.0 to -32.8), while for men, it decreased by 59.9% (-75.0 to -34.7). Additionally, the age-standardized rate of DALYs per 100,000 population significantly decreased from 142.6 (111.9 to 181.2) in 1990 to 69.8 (55.7 to 88.5) in 2021, reflecting a reduction of 51.1% (-64.2 to -35.8). The reductions in age-standardized rates of DALYs per 100,000 population were 52.8% (-69.1 to -31.8) for women and 48.9% (-63.7 to -28.3) for men. Throughout the years from 1990 to 2021, women consistently exhibited higher ASPR, ASIR, age-standardized rates of DALYs, and ASDR per 100,000 population compared to men. The prevalence rates per 100,000 population of RHD were highest in the 30–34 age group, while the 15–19 age group exhibited the highest incidence rate per 100,000 population. The greatest rates of death and DALYs per 100,000 population were observed in individuals aged 95 years and older. Among the provinces, Qom, Zanjan, Semnan, Bushehr, and Yazd demonstrated the lowest overall burden of RHD, whereas Sistan and Baluchistan, Chahar Mahaal and Bakhtiari, and Tehran exhibited the highest burden. The overall burden of RHD in Iran has decreased over the past three decades. However, this burden is not uniformly distributed across different sexes and age groups. Incidence and prevalence rates of RHD are notably higher among younger populations, whereas DALYs and deaths predominantly impact the elderly. Furthermore, women experience a greater burden of RHD compared to men. These findings emphasize the need for targeted healthcare policies to address resource allocation and interventions for RHD management, especially focusing on the vulnerable populations identified in this study.

This comprehensive mixed-methods study explores how disparities in resources shape science achievement among secondary school students, offering new insights into both systemic and individual factors that affect learning. The investigation draws on extensive quantitative data to demonstrate that students attending schools with abundant facilities, well-qualified teachers, and plentiful learning materials consistently achieve higher science outcomes than those in under-resourced settings. The study verifies a pronounced link between resource equity and academic success, emphasizing the key mediating role of student motivation. Constructs such as self-efficacy and intrinsic motivation exhibit strong, positive associations with science performance, confirming their importance in supporting achievement. Qualitative findings illuminate how shortages in infrastructure, instructional support, and curricular materials in low-resource schools undermine instructional quality and weaken student motivational beliefs, consequently limiting active engagement and perpetuating achievement inequality. The conceptual framework guiding this research integrates structural inequities with psychological dimensions and contextual moderators, such as socio-cultural environment and students’ subjective perceptions. This holistic model provides clarity on the multiple, intersecting channels through which both external and internal factors drive educational outcomes. Ultimately, the results highlight the necessity for comprehensive policy and pedagogical interventions. Effective strategies should address resource allocation and simultaneously foster motivational engagement through culturally responsive curriculum, enriched teacher training, and support from families and local communities. The implications extend to educational stakeholders and researchers, urging continued investigation of longitudinal and context-specific solutions for advancing equitable science learning.

This paper develops a dual-level framework addressing resource allocation and scheduling challenges in customer-driven production systems involving multi-mode projects with predefined release/due dates. At the tactical level, global resources undergo time-dependent allocation across projects, with transfer costs incurred during dynamic redistribution. Operational-level scheduling utilises these allocations for detailed sub-project execution. The objective minimises total costs encompassing transfer, idle, and indirect expenses. To streamline resource transfers, we introduce a novel ‘blocking’ strategy that partitions time-varying allocations into distinct ‘resource blocks’. We propose an adaptive large neighbourhood search (ALNS) and genetic algorithm (GA) featuring a ‘project macro-mode – activity sequence – activity mode’ hybrid encoding that integrates operational objectives within the tactical framework. Numerical studies confirm the superiority of ALNS, demonstrating it achieves 92% faster computation than CPLEX on small-scale instances while maintaining only a 0.47% optimality gap, reduces costs by up to 13.2% versus GA across 80 test instances, and shows particularly high efficacy in resource-constrained scenarios. The framework demonstrates significant potential for application in complex manufacturing environments like engineer-to-order and make-to-order production, directly contributing to reduced lead times, lower costs, and improved system throughput. Sensitivity analysis provides actionable insights on cost-drivers for production and project management practitioners.

Objective: This research seeks to examine the patterns for patients with different disorders in a more specialized clinical environment in order to address resource allocation, enhance efficiency of service delivery system and design relevant measures. Methods: The audit was carried out by observing clinical data of patients from April, 2023 to October, 2023 in Allama Iqbal Memorial Teaching Hospital Physiotherapy Department. Diagnoses were categorized in to five primary groups: musculo skeletal disorders; Orthopedic diseases Neurological diseases/ Cardiac diseases Surgical/Trauma and Other disorders. Results: Majority of the patients35.52% were suffering from Orthopaedic problems. However, the other diseases accounted 31.0% in this study. The patients of surgical procedures and trauma were counted to 7.4% respectively. The frequency of patients with neurological and Cardiac disorders were 5.4% and 0.3% respectively. Conclusion: In this study the majority of the patients were from the orthopaedic category while the patients of other discipline of medicine also seek help from physiotherapy department.

In times of high uncertainty, global recession, and lack of adequate funding for humanitarian needs, it is crucial to optimize humanitarian operations to achieve high levels of productivity and efficiency and to cut operational costs without compromising delivery of humanitarian aid to the most vulnerable population. One of the major contributors to cost of operations is transportation costs. In the humanitarian field, fleets are operated to deliver aid and assistance to the needy and most vulnerable population around the globe. These fleets are run in the form of field trips, which constitute a vehicle, a professional driver, humanitarian aid officers, and aid cargo. This study aims at optimizing resources and reducing costs of humanitarian field trips using an automated VBA program. The VBA program utilizes optimization algorithms such as compatibility and overlapping matrices to optimize field trip requisitions. The program also uses the Hungarian method to optimally solve the assignment matrices of vehicles-to-field trips and drivers-to-field trips. The outcome of the study is a user-friendly tool which comprises as inputs: vehicles data, drivers data, field trips history, and field trips requisition plan; and provides as outputs: the optimum list of combined field trips, optimum vehicles-to-field trips assignment based on minimum cost of transportation, and optimum drivers-to-field trips assignment based on the maximum productivity of drivers. The outcome of such a practical and utilizable tool goes beyond the humanitarian context and can be further tested and applied in fleet units of various industries. This study demonstrates the successful application of classical operations research techniques, specifically the assignment problem and the Hungarian method, to address resource allocation challenges in a real-world humanitarian logistics setting, providing a valuable bridge between theory and practice. In addition to mathematical modeling, the research leverages real operational data from a humanitarian fleet management system, validating the tool’s effectiveness on a large dataset of over 3,000 trips. The implementation also accounts for driver fatigue constraints and security conditions in trip combinability, offering a highly realistic and ethical optimization framework. Future extensions may include stochastic modeling, full route optimization, and integration with GIS systems for enhanced operational planning.

Hand hygiene compliance among nurses is a critical component in preventing healthcare-associated infections (HAIs), particularly in high-risk settings such as emergency departments (EDs). This study investigates the factors influencing hand hygiene adherence among nurses working in the emergency departments of tertiary hospitals in Peshawar, Pakistan. Employing a cross-sectional design, data were collected from 132 nurses through a structured questionnaire. Key findings indicate that while the majority of nurses are aware of WHO guidelines for hand hygiene, adherence is hindered by institutional, individual, and systemic challenges. Among the barriers, inadequate resources were identified as the most significant, reported by 45% of respondents, followed by insufficient training (32%), and heavy workload (21%). A considerable proportion (79%) of nurses reported not having received training on hand hygiene in the last six months, further highlighting a gap in ongoing professional development—additionally, 51% experienced skin irritation from hand hygiene products, indicating a need for hypoallergenic alternatives. The study also revealed discrepancies in compliance patterns, with 39% of nurses washing their hands fewer than five times during an 8-hour shift. Despite these challenges, 94% of respondents acknowledged the critical role of hand hygiene in preventing HAIs, underscoring the potential for improvement with targeted interventions. The research emphasizes the need for comprehensive strategies to address resource allocation, enhance training programs, and improve institutional policies for hand hygiene. Recommendations include integrating regular training sessions, increasing accessibility to hand hygiene resources, and fostering a culture of accountability within emergency departments. Addressing these challenges is imperative for reducing HAIs, improving patient outcomes, and promoting a safer healthcare environment. This study contributes to the growing body of evidence on hand hygiene practices in resource-constrained settings and highlights actionable steps for improving compliance among emergency department nurses in Pakistan.

The National Health Insurance (NHI) Act 20 of 2023 aims for universal health coverage. However, rehabilitation professions, especially physiotherapy, had limited involvement during key phases of NHI policy development, including the Green and White Papers, pilot projects, and the NHI Bill. President Cyril Ramaphosa enacted the NHI Bill in May 2024. To assess South African physiotherapists' perceptions and attitudes towards NHI, focusing on their perception of its objectives and implications for their profession. An online survey was conducted among 146 South African physiotherapists. A quantitative, non-experimental online survey was used. The data analysis revealed significant demographic influences on perceptions regarding NHI. Gender, age and professional experience played a role in shaping responses. Male physiotherapists were more likely than their female counterparts to perceive NHI as a means of addressing past healthcare disparities and increasing universal coverage. Professional experience and qualifications also played a crucial role, with distinct perspectives based on respondents' qualifications. Age influenced opinions on the impact of NHI on physiotherapists in private practice, with younger physiotherapists perceiving more negative impacts compared to older colleagues. Physiotherapists acknowledge NHI's potential to address healthcare disparities, but express concerns about its implementation and impact. They advocate for more inclusive policymaking, better communication, and improved strategies to ensure NHI meets diverse healthcare needs nationwide. Developing demographic-sensitive strategies and addressing resource allocation and infrastructure challenges are crucial to implementing NHI effectively.

In the global digitalization era, Internet+ technologies are reshaping physical education by integrating smart terminals, the Internet of Things, artificial intelligence, and cloud platforms to optimize teaching management and promote student health. This study illustrates how real-time physiological monitoring (e.g., heart rate, movement) and personalized training plans can address resource allocation gaps, track health behaviors, and support adaptive instruction. Empirical experiments conducted across multiple schools showed a 67.9% reduction in resource acquisition time, a 10.3% increase in lung capacity, and a 60.7% increase in post-class exercise duration. The platform's data-driven features—including safety alerts and tools for home–school collaboration—align with electronic and mobile health frameworks. Although the findings highlight the platform's potential, the limited geographic scope and small sample size warrant cautious interpretation. The discussion emphasizes the need for cross-regional validation, qualitative investigation of user experiences, and attention to ethical concerns related to data privacy and mental health.

This study investigated how proficiency in Somali, Arabic, and English predicts academic success of primary school students within Somaliland’s trilingual educational context. This research addresses a gap in large-scale, data-driven studies using advanced analytics. The study analyzed national examination data from 20,638 students who participated in the 2022/2023 Grade 8 national exams, sourced from the Somaliland National Examination and Certification Board (NECB). Methods included descriptive statistics, correlation analysis, multiple linear regression (MLR), and comparison of ten machine learning (ML) regression models—Linear, Polynomial, Robust, Partial Least Squares (PLS), Support Vector Regression (SVR), Principal Component Regression (PCR), Quantile, Ridge, Lasso, and Elastic Net Regression. Models were evaluated using Mean Absolute Percentage Error (MAPE), Root Mean Squared Percentage Error (RMSPE), Root Mean Squared Logarithmic Error (RMSLE), and Relative Root Squared Error (RRSE) to assess predictive accuracy. Findings showed that proficiency scores in Somali, Arabic, and English were significant positive predictors of overall academic performance, explaining 79.4% of the variance (R2 ≈ 0.794, F (3, 20,602) = 26,407.88, p < 0.001). English proficiency showed the strongest predictive coefficient (B = 2.34, p < 0.001), followed by Arabic (B = 2.23, p < 0.001), and Somali (B = 1.63, p < 0.001), highlighting their differential impact within the assessment framework. The ML model analysis revealed Polynomial Regression provided the most accurate predictions (lowest MAPE = 8.68%, lowest RRSE = 44.24%), suggesting non-linear relationships between language skills and academic achievement that linear models may not capture. The analysis revealed demographic imbalances, with data predominantly from urban (90.8%) and private school (57.3%) students. Policy implications emphasize enhancing equitable access to language instruction across all three languages, focusing on rural and public school populations; evaluating assessment practices for linguistic fairness; and addressing resource allocation disparities using ML insights for targeted interventions. Future research recommendations include longitudinal studies to explore causality, integrating comprehensive language assessments and socioeconomic data, investigating multilingual classroom practices, applying Explainable AI (XAI) techniques, examining language-demographic interactions, and analyzing subject-specific outcomes. Clinical Trial Registration: This study does not involve a clinical trial requiring registration.

ABSTRACT Project planning and scheduling is a critical stage in the construction industry that still faces significant challenges such as cost overruns and time delays, primarily due to the dominance of traditional network analysis techniques like the Critical Path Method (CPM), Program Evaluation and Review Technique (PERT), etc. Critical Chain Method emerged as an alternative, incorporating new tools to address the limitations of CPM such as path merging, availability of resources and extra safety time. Still, it has shortcomings in addressing resource allocation because there is no proven effective resource allocation algorithm. The current paper integrating the Earliest Late Start (ELS) technique with the critical chain approach for managing resource allocation on construction projects. Concise criteria were illustrated to deal with critical chain scheduling, including the implementation of the ELS resource allocation method. In the meantime, the explicit critical chain scheduling and the visual identification of the critical chain network using an Excel-based model are provided. The methodology of the modification emphasizes addressing the probabilities of the resource leveling calculations in the critical chain scheduling identified through the literature survey. The methodology is validated utilizing benchmark example from the literature and a survey was conducted and distributed among engineers, selected to investigate their feedback on the modified approach. Finally, the results show the potential application of the developed technique to reduce project duration and cost and monitor only one critical chain during the project period.

Abstract Efficient sequencing of processes and resource allocation are critical in production planning scenarios, such as manufacturing workshops and construction projects, to enhance efficiency and reduce operational costs. Resource allocation in such environments is often challenged by temporal constraints, process interdependencies, and resource limitations, which complicate scheduling and increase the risk of delays. This study presents a multi‐agent‐based simulation system to address these challenges. A scheduling optimisation model is developed to simulate and optimise resource allocation in complex processes with network structures and temporal constraints. The primary objective is to minimise production completion time while ensuring effective resource allocation. Additionally, an adaptive, partially distributed Agent‐Based Modelling and Simulation framework is proposed to simulate the execution logic of real‐world processes, integrating key factors such as resource limitations, process interdependencies, and real‐time decision‐making. A priority‐based genetic algorithm is also designed and embedded into the multi‐agent system to further optimise process sequencing and resource distribution. Simulation experiments across varying case scales validate the model and algorithm. This study highlights the potential of agent‐based simulation for solving complex engineering challenges and provides new insights for addressing resource allocation problems in network‐structured, time‐constrained environments.

Effective supply chain management is crucial for businesses to remain competitive in today's dynamic market. Despite extensive research, there is a lack of integrated approaches that simultaneously address resource allocation, routing, and delivery scheduling under uncertain conditions. This study develops a hybrid framework that combines robust optimization, simulated annealing, and reinforcement learning to enhance supply chain operations in complex networks involving fixed suppliers, distribution centers, and customers. Empirical results from rigorous testing demonstrate significant efficiency improvements and adaptability across diverse scenarios. A real-world case study from the logistics sector highlights the practical benefits, achieving notable cost savings and operational robustness. Sensitivity analysis further underscores the model’s capability to adapt to parameter variations. These findings emphasize the value of incorporating learning-based strategies into supply chain optimization, offering a powerful tool for organizations to address uncertainty and enhance decision-making efficiency.

This study examines the roles of educational planners in shaping pupils' academic performance in Dodoma Municipality, Tanzania. The study aims to address two main specific objectives namely identifying strategies employed for enhancing pupils’ academic performance and assessing the challenges faced by educational planners in achieving the goals. An interpretive paradigm guided the study, using a qualitative approach with an exploratory case study design. The target population included 2600 teachers and 37 educational officers in Dodoma municipality. However, from the total population of 2637 only 15 participants were purposively sampled to participate in the study. Data were analyzed using content analysis with the aid of Atlas.ti software; and ethical approval was granted by the Directorate of Research, Publications, and Innovation at the Open University of Tanzania. The findings revealed that the strategies employed by educational planners to improve pupils' academic performance include curriculum development and implementation, teacher training and professional development, resource allocation and infrastructure development, monitoring and evaluation, stakeholder engagement, policy evaluation and implementation, and promoting equity and inclusion. Despite these efforts, several challenges hinder the effectiveness of these strategies. These challenges include economic constraints, inadequate resources, demographic factors, socio-cultural influences, technological gaps, and the need for enhanced teacher capacity building. In conclusion, while educational planners in Dodoma Municipality have implemented strategies aimed at improving academic outcomes, significant challenges remain. Addressing these barriers is essential for sustainable improvement. The study recommends aligning the curriculum with local contexts, expanding teacher training opportunities, particularly in modern pedagogies, and increasing community involvement in educational planning. Additionally, it is crucial to address resource allocation and infrastructure disparities, particularly in rural areas, to ensure that all pupils benefit from improved educational opportunities.

This paper presents a novel algorithm to address resource allocation and network-slicing challenges in multiaccess edge computing (MEC) networks. Network slicing divides a physical network into virtual slices, each tailored to efficiently allocate resources and meet diverse service requirements. To maximize the completion rate of user-computing tasks within these slices, the problem is decomposed into two subproblems: efficient core-to-edge slicing (ECS) and autonomous resource slicing (ARS). ECS facilitates collaborative resource distribution through cooperation among edge servers, while ARS dynamically manages resources based on real-time network conditions. The proposed solution, a multiagent actor–critic resource scheduling (MAARS) algorithm, employs a reinforcement learning framework. Specifically, MAARS utilizes a multiagent deep deterministic policy gradient (MADDPG) for efficient resource distribution in ECS and a soft actor–critic (SAC) technique for robust real-time resource management in ARS. Simulation results demonstrate that MAARS outperforms benchmark algorithms, including heuristic-based, DQN-based, and A2C-based methods, in terms of task completion rates, resource utilization, and convergence speed. Thus, this study offers a scalable and efficient framework for resource optimization and network slicing in MEC networks, providing practical benefits for real-world deployments and setting a new performance benchmark in dynamic environments.

Restless multi-armed bandits (RMABs) have been widely utilized to address resource allocation problems with Markov reward processes (MRPs). Existing works often assume that the dynamics of MRPs are known prior, which makes the RMAB problem solvable from an optimization perspective. Nevertheless, an efficient learning-based solution for RMABs with unknown system dynamics remains an open problem. In this paper, we fill this gap by investigating a cooperative resource allocation problem with unknown system dynamics of MRPs. This problem can be modeled as a multi-agent online RMAB problem, where multiple agents collaboratively learn the system dynamics while maximizing their accumulated rewards. We devise a federated online RMAB framework to mitigate the communication overhead and data privacy issue by adopting the federated learning paradigm. Based on this framework, we put forth a Federated Thompson Sampling-enabled Whittle Index (FedTSWI) algorithm to solve this multi-agent online RMAB problem. The FedTSWI algorithm enjoys a high communication and computation efficiency, and a privacy guarantee. Moreover, we derive a regret upper bound for the FedTSWI algorithm. Finally, we demonstrate the effectiveness of the proposed algorithm on the case of online multi-user multi-channel access. Numerical results show that the proposed algorithm achieves a fast convergence rate of <inline-formula><tex-math notation="LaTeX">$\mathcal {O}(\sqrt{T\log (T)})$</tex-math></inline-formula> and better performance compared with baselines. More importantly, its sample complexity reduces sublinearly with the number of agents.

Mutual interference has been a critical issue in multiple unmanned aerial vehicles (multi-UAV) networks. As an advanced technology, digital twin (DT) maps physical entities into virtual domain, enables real-time monitoring and dynamic updates, thereby enhancing the adaptability and performance of multi-UAV networks. In this letter, we investigate joint spectrum allocation and power control for a multi-UAV radar sensing network, where multiple unmanned aerial vehicles (UAVs) simultaneously perform radar sensing separately to detect targets and avoid collision. By modeling the multi-UAV network as a graph, we employ graph neural network (GNN) to capture environmental features, construct the DT network, and address resource allocation issues. In particular, we propose a message-passing neural network based spectrum allocation method and a graph attention network based power control method to maximizing the minimum radar echo signal-to-interference-plus-noise ratio (SINR) among all UAVs. Simulation results show that the proposed DT-enhanced GNN based resource allocation method can significantly improve the minimum SINR and extend the sensing coverage.

This letter presents deep learning approaches for addressing resource allocation problems in wireless-powered communication networks. Conventional deep neural network (DNN) methods require the global channel state information (CSI), invoking impractical centralized operations. Also, their computations depend on the user population, which lacks the scalability of the network size. To this end, we propose decentralized and scalable DNN architectures. We interpret the ideal centralized DNN as a nomographic function that can be decomposed into multiple component DNNs. Each of these is dedicated to processing the local CSI of individual users, thereby leading to the decentralized architecture. To reduce coordination overheads among the H-AP and users, individual users leverage a DNN that compresses local CSI into low-dimensional messages shared with the H-AP. Since these DNN modules are designed to share identical trainable parameters, the proposed learning architecture can be universally applied to various configurations with arbitrary user populations. Numerical results show that the proposed decentralized method achieves almost identical performance to centralized schemes with reduced complexity.