Journal of Computational Science最新文献

Healthcare systems worldwide are increasingly subject to in-depth analysis. Problems in healthcare systems are of concern to the general public. For example, overcrowding in emergency departments creates several issues including longer waiting times, more frequent medical errors, a longer length of stay and worsened performance indicators. Overcrowding situations reduce the availability of staff and material resources, and therefore deteriorate the quality of care. The main cause of the overcrowding in emergency departments is the permanent interferences between the scheduled patients, unscheduled patients and urgent and unscheduled patients arriving at the emergency department. The objective of the present study is to develop an innovative decision support system that minimizes these interferences, while taking into account the perturbations that can occur throughout the day. The research’s ultimate goal is to improve the performance indicators via two processes: the first is a memetic algorithm based on a four dimensional hypercube genetic algorithm and local search techniques, and the second is based on a multi-agent system which dynamically orchestrates the patient pathway (given by the scheduling algorithm). In order to test and validate our approach, experiments are designed with real data from the adult emergency department at Lille University Medical Center. Simulations showed that with our approach we were able to reduce the waiting time of patients by 28.12%.

The brain is a complex system with functional and structural networks. Different neuroimaging methods have been developed to explore these networks, but each method has its own unique strengths and limitations, depending on the signals they measure. Combining techniques like electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) has gained interest, but understanding how the information derived from these modalities is related to each other remains an exciting open question. The multilayer network model has emerged as a promising approach to integrate different sources data. In this study, we investigated the hemodynamic and electrophysiological data captured by fNIRS and EEG to compare brain network topologies derived from each modality, examining how these topologies vary between resting state (RS) and task-related conditions. Additionally, we adopted the multilayer network model to integrate EEG and fNIRS data and evaluate the benefits of combining multiple modalities compared to using a single modality in capturing characteristic brain functioning.

A small-world network structure was observed in the rest, right motor imagery, and left motor imagery tasks in both modalities. We found that EEG captures faster changes in neural activity, thus providing a more precise estimation of the timing of information transfer between brain regions in RS. fNIRS provides insights into the slower hemodynamic responses associated with longer-lasting and sustained neural processes in cognitive tasks. The multilayer approach outperformed unimodal analyses, offering a richer understanding of brain function. Complementarity between EEG and fNIRS was observed, particularly during tasks, as well as a certain level of redundancy and complementarity between the multimodal and the unimodal approach, which depends on the modality and the specific brain state. Overall, the results highlight differences in how EEG and fNIRS capture brain network topology in RS and tasks and emphasize the value of integrating multiple modalities for a comprehensive view of brain connectivity and function.

Support vector machines (SVMs) are classic binary classification algorithms and have been shown to be a robust and well-behaved technique for classification in many real-world problems. However, there are ambiguities in the basic concepts of SVMs although these ambiguities do not affect the effectiveness of SVMs. Corinna Cortes and Vladimir Vapnik, who presented SVMs in 1995, pointed out that an SVM predicts through a hyperplane with a maximal margin. However existing literatures have two different definitions of the margin. On the other hand, Corinna Cortes and Vladimir Vapnik converted an SVM into an optimization problem that is much easier to solve. Nevertheless, existing papers do not explain how the optimization problem derives from an SVM well. These ambiguities may cause certain troubles in understanding the basic concepts of SVMs. For this purpose, this paper defines a separating hyperplane of a training data set and, hence, an optimal separating hyperplane of the set. The two definitions are reasonable since this paper proves that ${\text{w}}_0^{\text{T}}\text{x}+b_0=0$ is an optimal separating hyperplane of a training data set when ${\text{w}}_0$ and $b_0$ constitute a solution to the above optimization problem. Some notes on the above margin and optimization problem are given based on the two definitions. These notes should be meaningful for clarifying the basic concepts of SVMs.

Advances in connected and autonomous vehicle (CAV) technologies have made signal-free intersections a viable option for enhancing traffic performance. In the absence of traffic signal control, sequencing control strategies become crucial to ensuring the safety and efficiency of conflicting traffic flows at these intersections. The First-Come-First-Serve (FCFS) and Longest-Queue-First (LQF) strategies have received significant attention as fundamental approaches to managing connected and automated vehicles at signal-free intersections, serving as baselines for evaluating innovative strategies. However, the impact of varying traffic demand in conflicting directions on the volatility of CAV queues at signal-free intersections remains unclear, and there is a lack of analytical quantitative estimates on how these two fundamental sequencing strategies affect fairness within CAV queues. Furthermore, in urban road networks, CAVs entering a downstream intersection typically originate from an upstream intersection, and thus CAVs typically move in bunching and correlation. However, this phenomenon has received little attention in the modelling of CAV queues. To this end, in this paper, by virtue of the salient advantage of the Markovian Arrival Process (MAP) in describing the bunching and correlated arrival properties, an MAP-based double-input queueing model and its computational framework are developed to estimate the queueing process of CAVs at signal-free intersections. Some basic statistical metrics, such as queue length, delay, conditional queue length, and queue length variance, are derived. Additionally, numerical experiments are conducted to examine the queueing performance of FCFS and LQF strategies under different traffic conditions. The results suggest that the effectiveness of FCFS and LQF strategies varies depending on the level of traffic demand in the conflicting directions.

Hardness is a materials’ property with implications in several industrial fields, including oil and gas, manufacturing, and others. However, the relationship between this macroscale property and atomic (i.e., microscale) properties is unknown and in the last decade several models have unsuccessfully tried to correlate them in a wide range of chemical space. The understanding of such relationship is of fundamental importance for discovery of harder materials with specific characteristics to be employed in a wide range of fields. In this work, we have found a physical descriptor for Vickers hardness using a symbolic-regression artificial-intelligence approach based on compressed sensing. SISSO (Sure Independence Screening plus Sparsifying Operator) is an artificial-intelligence algorithm used for discovering simple and interpretable predictive models. It performs feature selection from up to billions of candidates obtained from several primary features by applying a set of mathematical operators. The resulting sparse SISSO model accurately describes the target property (i.e., Vickers hardness) with minimal complexity. We have considered the experimental values of hardness for binary, ternary, and quaternary transition-metal borides, carbides, nitrides, carbonitrides, carboborides, and boronitrides of 61 materials, on which the fitting was performed.. The found descriptor is a non-linear function of the microscopic properties, with the most significant contribution being from a combination of Voigt-averaged bulk modulus, Poisson’s ratio, and Reuss-averaged shear modulus. Results of high-throughput screening of 635 candidate materials using the found descriptor suggest the enhancement of material’s hardness through mixing with harder yet metastable structures (e.g., metastable VN, TaN, ReN${}_2$, Cr${}_3$N${}_4$, and ZrB${}_6$ all exhibit high hardness).

Automated code generation allows for a separation between the development of a model, expressed via a domain specific language, and lower level implementation details. Algorithmic differentiation can be applied symbolically at the level of the domain specific language, and the code generator reused to implement code required for an adjoint calculation. However the adjoint calculations are complicated by the well-known problem of storing or recomputing the forward data required by the adjoint, and different checkpointing strategies have been developed to tackle this problem. This article considers the combination of high-level algorithmic differentiation with step-based checkpointing schedules, with the primary application being for solvers of time-dependent partial differential equations. The focus is on algorithmic differentiation using a dynamically constructed record of forward operations, where the precise structure of the original forward calculation is unknown ahead-of-time. In addition, high-level approaches provide a simplified view of the model itself. This allows data required to restart and advance the forward, and data required to advance the adjoint, to be identified. The difference between the two types of data is here leveraged to implement checkpointing strategies with improved performance.

In the realm of fluid mechanics, where computationally-intensive simulations demand significant time investments, especially in predicting aerodynamic coefficients, the conventional use of time series forecasting techniques becomes paramount. Existing methods prove effective with periodic time series, yet the challenge escalates when faced with aperiodic or chaotic system responses. To address this challenge, we introduce DARSI (Deep Auto-Regressive Time Series Inference), an advanced architecture and an efficient hybrid of Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) components. Evaluated against established architectures (CNN, DLinear, LSTM, LSTNet, and PatchTST) for forecasting Coefficient of Lift ($C_L$) values corresponding to Angles of Attack (AoAs) across periodic, aperiodic, and chaotic regimes, DARSI demonstrates remarkable performance, showing an average increase of 79.95% in CORR, 76.57% reduction in MAPE, 94.70% reduction in MSE, 76.18% reduction in QL, and 75.21% reduction in RRSE. Particularly adept at predicting chaotic aerodynamic coefficients, DARSI emerges as the best in static scenarios, surpassing DLinear and providing heightened reliability. In dynamic scenarios, DLinear takes the lead, with DARSI securing the second position alongside PatchTST. Furthermore, static AoAs at 24.7 are identified as the most chaotic, surpassing those at 24.9 and the study reveals a potential inflection point at AoA 24.7 in static scenarios for both DLinear and DARSI, warranting further confirmation. This research positions DARSI as an adept alternative to simulations, offering computational efficiency with significant implications for diverse time series forecasting applications across industries, particularly in advancing aerodynamic predictions in chaotic scenarios.

The resonance of tsunami waves in semi-enclosed bays is paramount in understanding and mitigating the impact of seismic events on coastal communities. Semi-enclosed bays, characterized by their partial enclosure, can amplify the effects of incoming tsunami waves due to resonance behavior, where the natural frequencies of the bay correspond to those of the incoming waves. This resonance phenomenon can significantly increase wave height and inundation levels, posing an increased risk to nearby settlements and infrastructure. Understanding the resonance patterns in these bays is crucial for accurate hazard assessment, early warning systems, and effective disaster preparedness and response strategies. On October 30, 2020, an earthquake occurred between the Turkish Bay of Seferihisar Bay and the Greek island of Samos in the Aegean Sea. Long waves generated by the normal-faulting earthquake caused notable damage to settlements within Seferihisar Bay and the north coast of Samos Island. According to the measurements of the Syros mareograph stations, the wave heights were between 2 and 20 cm and wave periods between 9 and 20 seconds. Based on on-site survey reports conducted after the earthquake, inundation was reported in six settlements within Seferihisar Bay. However, inundation was notably higher in Sığacık and Akarca, reaching 2–3 times higher than in other locations, and the water level reached 2 m high. Given that the variance in inundation levels is attributed to resonance phenomena in Sığacık and Akarca rather than the propagation of tsunami waves, this study focused on conducting wave resonance modeling in Seferihisar Bay. The resonance modeling was performed using the RIDE wave model. Furthermore, the research has been expanded to assess the resonance patterns that might emerge in the event of an alternative earthquake or underwater landslide along the fault line responsible for the seismic event, encompassing wave periods ranging from T = 1–9 minutes and T = 20–30 minutes. Modeling results revealed that on the day of the earthquake, wave heights in Sığacık Marina and Akarca surged by 8.5 times in comparison to the wave height at the epicenter. This increase is notably higher, ranging from 2 to 2.5 times, compared to calculations made for other locations (Demircili, Altınköy, and Tepecik). Consequently, it was concluded that one of the reasons for the heightened effectiveness of inundation in Sığacık and Akarca was attributable to resonance. Moreover, supplementary investigations have indicated that waves with a period of T<9 minutes will pose higher risks for Demircili, Altınköy, Sığacık Marina, and Tepecik compared to the day of the earthquake. By comprehensively studying wave resonance in semi-enclosed bays, researchers and policymakers can better anticipate the potential impact of tsunami events and take measures to protect coastal communities, ultimately increasing resilience and reducing the loss of life and property in vulner