Search

Food allergy is a significant public health concern and food safety issue. Deriving from classical toxicology principle, the food allergen risk assessment has been considered a science‐based strategy to identify, quantify, and manage the food allergy risks as such risk represent a significant food safety. Moreover, the implication of the precautionary allergen labeling in most jurisdictions is voluntary, resulting potential risk to allergic consumers. In this study, a quantitative risk assessment technique was employed to evaluate the risk of wheat allergy in prepackaged foods. The assessment utilized probabilistic models, including the lognormal, Weibull, Gamma distributions, and Bayesian model averaging. The predicted allergic reactions were determined to be 682, 854, 677, and 721 incidents per 10,000 eating occasions within wheat allergic population, respectively. The findings of this study revealed that the consumption of prepackaged foods containing gluten without wheat/gluten summary (i.e., ingredient) labeling would potentially pose the risk of allergic reactions to wheat allergic individuals. The utilization of quantitative risk assessment methodology at different points of the food system facilitates alleviating unnecessary concerns to stakeholders while maintaining a reasonable knowledge of allergy risk and providing valuable guidance in formulating effective management strategies to mitigate the risk of food allergies, thereby contributing to the overall safety of the sustainable food system. [ABSTRACT FROM AUTHOR]

Conceptual pre-estimation of road tunnel costs is a vital yet challenging process during feasibility studies due to the prevailing underground uncertainties and risks. Most existing cost estimation approaches are deterministic, probabilistic methods are limited to a single cost examination. This paper introduces a probabilistic model for pre-estimating road tunnels life-cycle costs. It aims to capture their inherent uncertainty holistically, thus enabling more reliable decision-making at a project's early stages. The proposed model is developed in three steps: multiple regression analysis, fitting distribution, and Monte-Carlo simulation. The first step unveils the correlations between independent and dependent variables, while the other two steps return the probabilistic descriptions of cost drivers and life-cycle costs. Civil engineering, electromechanical works costs, energy consumption, and operation & maintenance costs are determined as the model's output variables. A real-world database of 32 dual-bore road tunnels with a total length of 55 km is used to develop probability distributions within reasonable confidence intervals and a sensitivity analysis was performed utilizing tornado charts. The results can assist public authorities and managers in estimating the probability of a specific life-cycle cost and making appropriate decisions in accordance with their risk tolerance. [ABSTRACT FROM AUTHOR]

This paper presents a probabilistic model for various machine learning (ML) applications. While deep learning (DL) has produced state-of-the-art results in many domains, DL models are complex and over-parameterized, which leads to high uncertainty about what the model has learned, as well as its decision process. Further, DL models are not probabilistic, making reasoning about their output challenging. In contrast, the proposed model, referred to as Yet Another Discriminate Analysis(YADA), is less complex than other methods, is based on a mathematically rigorous foundation, and can be utilized for a wide variety of ML tasks including classification, explainability, and uncertainty quantification. YADA is thus competitive in most cases with many state-of-the-art DL models. Ideally, a probabilistic model would represent the full joint probability distribution of its features, but doing so is often computationally expensive and intractable. Hence, many probabilistic models assume that the features are either normally distributed, mutually independent, or both, which can severely limit their performance. YADA is an intermediate model that (1) captures the marginal distributions of each variable and the pairwise correlations between variables and (2) explicitly maps features to the space of multivariate Gaussian variables. Numerous mathematical properties of the YADA model can be derived, thereby improving the theoretic underpinnings of ML. Validation of the model can be statistically verified on new or held-out data using native properties of YADA. However, there are some engineering and practical challenges that we enumerate to make YADA more useful. [ABSTRACT FROM AUTHOR]

The growth optimizer (GO) is an innovative and robust metaheuristic optimization algorithm designed to simulate the learning and reflective processes experienced by individuals as they mature within the social environment. However, the original GO algorithm is constrained by two significant limitations: slow convergence and high memory requirements. This restricts its application to large-scale and complex problems. To address these problems, this paper proposes an innovative enhanced growth optimizer (eGO). In contrast to conventional population-based optimization algorithms, the eGO algorithm utilizes a probabilistic model, designated as the virtual population, which is capable of accurately replicating the behavior of actual populations while simultaneously reducing memory consumption. Furthermore, this paper introduces the Lévy flight mechanism, which enhances the diversity and flexibility of the search process, thus further improving the algorithm's global search capability and convergence speed. To verify the effectiveness of the eGO algorithm, a series of experiments were conducted using the CEC2014 and CEC2017 test sets. The results demonstrate that the eGO algorithm outperforms the original GO algorithm and other compact algorithms regarding memory usage and convergence speed, thus exhibiting powerful optimization capabilities. Finally, the eGO algorithm was applied to image fusion. Through a comparative analysis with the existing PSO and GO algorithms and other compact algorithms, the eGO algorithm demonstrates superior performance in image fusion. [ABSTRACT FROM AUTHOR]

Unobserved heterogeneity of crashes remains a significant issue for freeways that influence crash prediction, and therefore deserves much attention. Using a fusion data set of crash data, driving behavior data, and traffic flow data, this study explores the spatiotemporal heterogeneity of crash determinants for different freeway segments (e.g. Yixing section and Liyang section of Ning-Hang freeway of China) and then predict the crash probability. A random effect negative binomial regression model is built to investigate the contributing factors of the crashes. Remarkable differences are observed in the crash determinants for Yixing section (include average vehicle speed, hourly average traffic volume, average free speed, road segment length, and number of left lane-merging) and Liyang section (include average intensity of aggressive driving behavior, average kilometer traffic volume). The results found the traffic flow has a more significant impact on crashes than the driving behaviors. It is found that the crash probability is a monotone decreasing function when the predicted number of crash is 0. With the increase of the number of predicted crash, the crash probability gradually converges from a large value to 0. Then the probability of other predicted number of crashes (e.g. crash = 1, crash = 2, crash = 3) presents a quadratic parabola trends. The model comparison demonstrates that the proposed model outperforms conventional model, and the prediction performance for Liyang section is better than that of Yixing section. The research findings are interesting and important for preventing crashes. [ABSTRACT FROM AUTHOR]

Prediction of flood hydrographs in semiarid regions is a complex task due to limited rainfall-runoff observations. The application of complex loss models that have intensive input data requirements can be impractical for such regions. The performance of three loss models, namely, initial and constant rate (IC), Soil Conservation Service (SCS), and constant fraction (CF), in prediction of flood events in a 37.2-km2 semiarid watershed was evaluated using deterministic and probabilistic hydrologic models. We quantified the performance in terms of bias, error, and correlation via relative error (RE), Nash–Sutcliff efficiency (NSE), and percent bias (PBIAS) for 14 events with dry prestorm conditions and a range of rainfall properties (duration, depth, and temporal pattern) and runoff characteristics (peak, volume, and time to peak). The NSE values of the deterministic model ranged from 0.61 to 0.90 and −0.50 to 0.63 for calibration and validation, respectively, in the best model (IC). The results suggest that the performance of loss models was inconsistent in terms of hydrograph attributes. The IC model was best in terms of peak flow according to both deterministic and probabilistic models and best in terms of volume according to the deterministic model, but similar to SCS and better than the CF based on the probabilistic model. The CF model mostly underestimated the runoff volume and peak flow. The performances of the loss models were almost identical in the prediction of the time to peak. These results suggested that deterministic models may be insufficient for selecting the best loss models. Probabilistic models, incorporating the parametric uncertainty, are needed to further evaluate the performance of loss models. There was no correlation between the performance of models and the size of events. Rainfall temporal pattern was found to be an effective factor in the accuracy of flood hydrology predictions. The results can guide the selection of loss models in semiarid watersheds. [ABSTRACT FROM AUTHOR]

Reconstruction of data loss in structural seismic responses is important for structural health monitoring to evaluate the safety of structures. A physics-guided neural network that leverages the prior knowledge was proposed for reconstructing structural seismic responses that were inaccessible to measure or missing during earthquakes. The presented methodology consisted of convolutional neural networks with dilated kernel and fully connected neural networks, which were developed to achieve a multitask learning that involved the regression task with measured labeled displacement data and the reconstruction task of seismic response without any labels. To better balance the loss gradient across different tasks, a probabilistic model was introduced to optimize the weight coefficient for each task by quantifying the task-dependent uncertainty based on Bayesian statistics. The weight coefficient for each task can be dynamically updated during the training process, thereby improving the learning efficacy and performance accuracy of the neural networks. The probabilistic model with task-dependent uncertainty was validated to outperform the equal-weighted model (i.e. equal weight for each task) in reconstructing the structural seismic responses based on numerical data, even when the relevant physical information (i.e. Bouc–Wen model) was not complete. [ABSTRACT FROM AUTHOR]

Model-based hyperspectral image (HSI) denoising methods have attracted continuous attention in the past decades, due to their effectiveness and interpretability. In this work, we aim at advancing model-based HSI denoising, through sophisticated investigation for both the fidelity and regularization terms, or correspondingly noise and prior, by virtue of several recently developed techniques. Specifically, we formulate a novel unified probabilistic model for the HSI denoising task, within which the noise is assumed as pixel-wise non-independent and identically distributed (non-i.i.d) Gaussian predicted by a pre-trained neural network, and the prior for the HSI image is designed by incorporating the deep image prior (DIP) with total variation (TV) and spatio-spectral TV. To solve the resulted maximum a posteriori (MAP) estimation problem, we design a Monte Carlo Expectation–Maximization (MCEM) algorithm, in which the stochastic gradient Langevin dynamics (SGLD) method is used for computing the E-step, and the alternative direction method of multipliers (ADMM) is adopted for solving the optimization in the M-step. Experiments on both synthetic and real noisy HSI datasets have been conducted to verify the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Cell trajectory inference is very important to understand the details of tissue cell development, state differentiation and gene dynamic regulation. However, due to the high noise and heterogeneity of the single-cell data, it is challenging to infer cell trajectory in complex biological processes. Here, we proposed a new trajectory inference method, called Metric Learning Bhattacharyya Kernel Feature Decomposition (MLBKFD). In MLBKFD, a statistical model was used to infer cell trajectory by calculating the Markov transition matrix and Bhattacharyya kernel matrix between cells. Before that, to expedite the matrix calculation in the statistical model, a deep feedforward neural network was used to perform dimensionality reduction on single-cell data. The MLBKFD was evaluated on four typical datasets as well as seven recent human fetal lung datasets. Comparisons with the two outstanding methods (i.e., DTFLOW and MARGARET) demonstrate that the MLBKFD is capable of accurately inferring cell development and differentiation trajectories from single-cell data with different sizes and sources. Notably, MLBKFD exhibits nearly twice the speed of DTFLOW while maintaining high precision, particularly when dealing with large datasets. MLBKFD provides accurate and efficient trajectory inference, empowering researchers to gain deeper insights into the complex dynamics of cell development and differentiation. 1. MLBKFD is a powerful probabilistic model to effectively infer cell pseudo trajectories from single-cell data by calculating the Markov transition matrix and Bhattacharyya kernel matrix between cells. 2. MLBKFD exhibits fast speed and excellent visualization effects on cell trajectory inference, by utilizing a deep feedforward neural networks for dimensionality reduction of single-cell data. 3. By iteratively refining cell labels in the deep neural networks, MLBKFD can achieve excellent cell clustering results, enabling it to infer cell trajectories without relying on true cell labels. [ABSTRACT FROM AUTHOR]

Pesticides and veterinary drugs are widely employed to support food production. Assessing potential risks associated with the dietary consumption of pesticide and veterinary drug residues is, however, essential. Potential risks depend on the toxicity degree of the analyzed residue and population exposure levels. Human populations are exposed to numerous chemical substances through different pathways with varying exposure times, leading to increased health risks when compared to exposure to individual chemicals. Cumulative exposure assessments usually assess combined exposures to multiple chemicals through multiple exposure pathways. In this sense, this comprehensive review aims to provide insights into cumulative dietary pesticide and veterinary drug residue exposures. The main methodologies, strategies, and legislation employed by international agencies to this end are discussed. A review concerning articles that apply existing methodologies and approaches, as well as the challenges in this context faced by Brazil is also presented. As this is a critical issue not only for Brazilian public health but also for the global community, regulatory agencies should prioritize formulating regulations that incorporate exposure assessments regarding the simultaneous presence of residues and contaminants in foodstuffs. [ABSTRACT FROM AUTHOR]

Cutting tool manufacturers face a tough challenge in developing custom solutions for specific customer requirements. Several trials are required, encompassing the selection of materials, tool configuration parameters, manufacturing of tools and testing under target conditions to arrive at an acceptable solution. In this work, the authors present a recommender system that utilizes a hierarchical deep learning-based machine learning model, handcrafted using domain knowledge, to predict top N tool configurations for a given target requirement with a probability score. The authors also discuss methods for data augmentation to deal with limited data as well as a probabilistic approach to predict the top N tool configurations from the trained models. The proposed system is applied to a case of centerless cylindrical grinding wheel selection problem. The outcomes indicate an overall accuracy of 92.4% for single best-fit specification, with 100% within the top five recommendations for past designs. Some of the alternatives proposed by the model are observed to be potentially superior to what was chosen earlier by experts. Without the selection hierarchy, a deep learning model achieved a single best-fit accuracy of 83.5% and the probabilistic model achieved a top-five recommendation accuracy of 89.8%, highlighting the merit of the hierarchical approach. [ABSTRACT FROM AUTHOR]

This paper presents a novel probabilistic data-driven approach to trip-level energy consumption estimation of battery electric vehicles (BEVs). As there are very few electric vehicle (EV) charging stations, EV trip energy consumption estimation can make EV routing and charging planning easier for drivers. In this research article, we propose a new driver behavior-centric EV energy consumption estimation model using probabilistic neural networks with model uncertainty. By incorporating model uncertainty into neural networks, we have created an ensemble of neural networks using Monte Carlo approximation. Our method comprehensively considers various vehicle dynamics, driver behavior, and environmental factors to estimate EV energy consumption for a given trip. We propose relative positive acceleration (RPA), average acceleration, and average deceleration as driver behavior factors in EV energy consumption estimation, and this paper shows that the use of these driver behavior features improves the accuracy of the EV energy consumption model significantly. Instead of predicting a single-point estimate for EV trip energy consumption, this proposed method predicts a probability distribution for the EV trip energy consumption. The experimental results of our approach show that our proposed probabilistic neural network with weight uncertainty achieves a mean absolute percentage error of 9.3% and outperforms other existing EV energy consumption models in terms of accuracy. [ABSTRACT FROM AUTHOR]

Abstract This study uses real and simulated information from 885 bridges in Brazil. A total of 2,655 available inspection data were collected from the database, and 37,170 additional data were simulated from deterministic deterioration prediction models developed in previous studies. The probabilistic Markov matrices-based models obtained include one covering all the bridges, specific models for non-aggressive and aggressive environments, and models for Average Daily Traffic (ADT) of less than and more than 4,000. Validation showed good metrics, with a coefficient of determination of 0.6268, a mean absolute error and mean squared error below 0.5, and an accuracy of 66.25%. Finally, these tools enable more accurate forecasting, and a better understanding of the risks associated with the deterioration of structures for safe and cost-effective bridge management.

Phytoextraction, utilizing plants to remove soil contaminants, is a promising approach for environmental remediation but its application is often limited due to the long time requirements. This study aims to develop simplified and user-friendly probabilistic models to estimate the time required for phytoextraction of contaminants while considering uncertainties. More specifically we: i) developed probabilistic models for time estimation, ii) applied these models using site-specific data from a field experiment testing pumpkin (Cucurbita pepo ssp. pepo cv. Howden) for phytoextraction of DDT and its metabolites (ΣDDX), iii) compared timeframes derived from site-specific data with literature-derived estimates, and iv) investigated model sensitivity and uncertainties through various modelling scenarios. The models indicate that phytoextraction with pumpkin to reduce the initial total concentration of ΣDDX in the soil (10 mg/kg dw) to acceptable levels (1 mg/kg dw) at the test site is infeasible within a reasonable timeframe, with time estimates ranging from 48–123 years based on literature data or 3 570–9 120 years with site-specific data using the linear or first-order exponential model, respectively. Our results suggest that phytoextraction may only be feasible at lower initial ΣDDX concentrations (< 5 mg/kg dw) for soil polishing and that alternative phytomanagement strategies should be considered for this test site to manage the bioavailable fraction of DDX in the soil. The simplified modes presented can be useful tools in the communication with site owners and stakeholders about time approximations for planning phytoextraction interventions, thereby improving the decision basis for phytomanagement of contaminated sites. [ABSTRACT FROM AUTHOR]

Nowadays, it is difficult to imagine a powerful algorithm of cryptography that can continue cryptanalyzing and attacking without the use of unconventional techniques. Although some of the substitution algorithms are old, such as Vigenere, Alberti, and Trithemius ciphers, they are considered powerful and cannot be broken. In this paper we produce the novelty algorithm, by using of biological computation as an unconventional search tool combined with an uninhibited analysis method is the vertical probabilistic model, that makes attacking and analyzing these ciphers possible and very easy to transform the problem from a complex to a linear one, which is a novelty achievement. The letters of the encoded message are processed in the form of segments of equal length, to report the available hardware components. Each letter codon represents a region of the memory strand, and the letters calculated for it are symbolized within the probabilistic model so that each pair has a triple encoding: the first is given as a memory strand encoding and the others are its complement in the sticker encoding; These encodings differ from one region to another. The solution space is calculated and then the parallel search process begins. Some memory complexities are excluded even though they are within the solution paths formed, because the natural language does not contain its sequences. The precision of the solution and the time consuming of access to it depend on the length of the processed text, and the precision of the solution is often inversely proportional to the speed of access to it. As an average of the time spent to reach the solution, a text with a length of 200 cipher characters needs approximately 15 minutes to give 98% of the correct components of the specific hardware. The aim of the paper is to transform OTP substitution analysis from a NP problem to a O(nm) problem, which makes it easier to find solutions to it easily with the available capabilities and to develop methods that are harnessed to attack difficult and powerful ciphers that differ in class and type from the OTP polyalphabetic substitution ciphers. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Railway Science & Engineering is the property of Journal of Railway Science & Engineering Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

A set of n drones with limited communication range is deployed to monitor a terrain partitioned into n pairwise disjoint and closed convex trajectories, one per drone. There is exactly one communication link between two trajectories if they are close enough, and drones can communicate provided they visit the link at the same time. If each robot flies around an assigned area and shares information with the neighbors periodically the system is said to be synchronized. Over time, one or more drones may fail and the ability to survey, communicate, and stay connected decreases, thus the robustness against drone failure becomes crucial. In this paper we study various problems related to the proper functioning of a synchronized system under drone failure. First, we provide efficient algorithms, both centralized and decentralized, for determining the connected components induced by the set of surviving drones. Second, we study coverage, isolation, and connectivity under a probabilistic failure model and show that, in the case of grids, the system is quite robust in the sense that it can tolerate a large probability of failure before drones fail to completely cover the terrain, become isolated, or the system loses full connectivity. [ABSTRACT FROM AUTHOR]

D2D edge computing is a promising solution to address the conflict between limited network capacity and increasing application demands, where mobile devices can offload their tasks to other peer devices/servers for better performance. Task offloading is critical to the performance of D2D edge computing. Most existing works on task offloading assume the task processing time is known or can be accurately estimated. However, the processing time is often uncertain until it is finished. Moreover, the same task can have largely different execution times under different scenarios, which leads to inaccurate offloading decisions and degraded performance. To address this problem, we propose a game-based probabilistic task offloading scheme with an uncertain processing time in D2D edge networks. First, we characterize the uncertainty of the task processing time using a probabilistic model. Second, we incorporate the proposed probabilistic model into an offloading decision game. We also analyze the structural properties of the game and prove that it can reach a Nash equilibrium. We evaluate the proposed work using real-world applications and datasets. The experimental results show that the proposed probabilistic model can accurately characterize the uncertainty of completion time, and the offloading algorithm can effectively improve the overall task completion rate in D2D networks. [ABSTRACT FROM AUTHOR]

Variations in the tumor genome can result in allelic changes compared to the reference profile of its homogenous body source on genetic markers. This brings a challenge to source identification of tumor samples, such as clinically collected pathological paraffin-embedded tissue and sections. In this study, a probabilistic model was developed for calculating likelihood ratio (LR) to tackle this issue, which utilizes short tandem repeat (STR) genotyping data. The core of the model is to consider tumor tissue as a mixture of normal and tumor cells and introduce the incidence of STR variants (φ) and the percentage of normal cells (Mxn) as a priori parameters when performing calculations. The relationship between LR values and φ or Mxn was also investigated. Analysis of tumor samples and reference blood samples from 17 colorectal cancer patients showed that all samples had Log10(LR) values greater than 1014. In the non-contributor test, 99.9% of the quartiles had Log10(LR) values less than 0. When the defense's hypothesis took into account the possibility that the tumor samples came from the patient's relatives, LR greater than 0 was still obtained. Furthermore, this study revealed that LR values increased with decreasing φ and increasing Mxn. Finally, LR interval value was provided for each tumor sample by considering the confidence interval of Mxn. The probabilistic model proposed in this paper could deal with the possibility of tumor allele variability and offers an evaluation of the strength of evidence for determining tumor origin in clinical practice and forensic identification. [ABSTRACT FROM AUTHOR]

The numerous factors that influence the technological process and cause errors during manufacture of parts complicate the problem of ensuring the required accuracy of required dimensions. The article proposes a method for determining coefficients of simulation models using probabilistic statistical methods, the versatility of which makes them applicable to a wide range of problems. [ABSTRACT FROM AUTHOR]

In the competitive market, process monitoring can ensure the quality of products, but strong nonlinearities, slow dynamics, and uncertainties characterize the complexities of the large-scale chemical plant. When the fault occurs, it will not influence the process instantaneously but will react after a few time points. After all the products affected by the faults are inspected, it is too late to fix the process. Conventional approaches neither do nor care about early detection before any disturbance significantly affects the process. To estimate disturbances propagated through the process, a multi-step prediction model is essential. The purpose of early process monitoring is to detect any problem with the currently running process as early as possible. In this paper, a multi-step prediction system is proposed. The system is a dynamic model that can capture the dynamic relationship of past process input variables and future process output variables. It provides a lower dimension and a lower noise-contaminated space for data analysis. Particularly, the past input and output process data can be mapped from the observation space into the latent space to acquire their intrinsic properties. The latent variables preserve the dynamic information for future multi-step prediction so that early warning can be achieved. An industrial example of the PVC dying process is presented to show the multistep predictive ability of the proposed method. [ABSTRACT FROM AUTHOR]

Traffic on highways has increased significantly in the past few years. Consequently, this has caused delays for the drivers in reaching their final destination and increased the highway's congestion level. Many options have been proposed to ease these issues. In this paper, a model of the highway drivers' population was built based on several factors, including the behavioral patterns of the drivers, like drivers' time flexibility to reach the destination, their carpool eligibility, and their tolerance to pay the toll price, in addition to the traffic information from the system. A fuzzy logic decision-making model is presented to emulate how drivers would choose the lane to use based on the aforementioned factors and the current congestion levels of all the lanes on the highway. The presented model, along with the simulation results from applying the model to different simulation scenarios, show the usefulness of such a model in predicting an optimal toll value. Such optimal value would reduce congestion on the highway at one end while maximizing the revenue for the toll company. [ABSTRACT FROM AUTHOR]

As immunological and clinical studies become more complex, there is an increasing need to analyze temporal immunophenotypes alongside demographic and clinical covariates, where each subject receives matrix-valued time series observations for potentially high-dimensional longitudinal features, as well as other static characterizations. Researchers aim to find the low-dimensional embedding of subjects using matrix-valued time series observations and investigate relationships between static clinical responses and the embedding. However, constructing these embeddings can be challenging due to high dimensionality, sparsity, and irregularity in sample collection over time. In addition, the incorporation of static auxiliary covariates is frequently desired during such a construction. To address these issues, we propose a smoothed probabilistic PARAFAC model with covariates (SPACO) that uses auxiliary covariates of interest. We provide extensive simulations to test different aspects of SPACO and demonstrate its application to an immunological dataset from patients with SARS-CoV-2 infection. for this article are available online. [ABSTRACT FROM AUTHOR]

Buildings have great energy flexibility potential to manage supply-demand imbalance in power grids with high renewable penetration. Accurate and real-time quantification of building energy flexibility is essential not only for engaging buildings in electricity and grid service markets, but also for ensuring the reliable and optimal operation of power grids. This paper proposes a probabilistic model for rapidly quantifying the aggregated flexibility of buildings under uncertainties. An explicit equation is derived as the analytical solution of a commonly used second-order building thermodynamic model to quantify the flexibility of individual buildings, eliminating the need of time-consuming iterative and finite difference computations. A sampling-based uncertainty analysis is performed to obtain the distribution of aggregated building flexibility, considering major uncertainties comprehensively. Validation tests are conducted using 150 commercial buildings in Hong Kong. The results show that the proposed model not only quantifies the aggregated flexibility with high accuracy, but also dramatically reduces the computation time from 3605 s to 6.7 s, about 537 times faster than the existing probabilistic model solved numerically. Moreover, the proposed model is 8 times faster than the archetype-based model and achieves significantly higher accuracy.

Airport tower control plays an instrumental role in ensuring airport safety. However, obtaining objective, quantitative safety evaluations is challenging due to the unavailability of pertinent human operation data. This study introduces a probabilistic model that combines aircraft dynamics and the peak-over-threshold (POT) approach to assess the safety performance of airport controllers. We applied the POT approach to model reaction times extracted from a radiotelephony dataset via a voice event detection algorithm. The model couples the risks of tower control and aircraft operation to analyze the influence of human factors. Using data from radiotelephony communications and the Base of Aircraft Data (BADA) database, we compared risk levels across scenarios. Our findings revealed heightened airport control risks under low demand (0.374) compared to typical conditions (0.197). Furthermore, the risks associated with coupling under low demand exceeded those under typical demand, with the final approach stage presenting the highest risk (4.929×10−7). Our model underscores the significance of human factors and the implications of mental disconnects between pilots and controllers for safety risks. Collectively, these consistent findings affirm the reliability of our probabilistic model as an evaluative tool for evaluating the safety performance of airport tower controllers. The results also illuminate the path toward quantitative real-time safety evaluations for airport controllers within the industry. We recommend that airport regulators focus on the performance of airport controllers, particularly during the final approach stage.

Various computing approaches within vehicular networks, such as vehicular edge computing (VEC) and cloud computing, have been suggested to facilitate task offloading, aiming to improve user satisfaction. The features of vehicular networks, including the rapid movement of vehicles and the fluctuating distribution of vehicle densities, present challenges to task offloading with in the VEC. Numerous algorithms have been suggested to address these challenges and provide an effective task-offloading framework. This paper introduces a probabilistic model that analyzes task offloading across different computing tiers, alongside proposing a mobile computing paradigm tailored to the dynamic nature of vehicular networks (VN). This paradigm aims to maintain persistent connectivity and enhanced connection reliability despite mobility facilitating sustainable end-to-end service delivery. Building upon this premise, we propose a three-tier computing paradigm comprising Vehicle Edge of Things (VEoTC), VEC, and Cloud Computing (CC). Within the VEoTC tier, Service Vehicles (SV) equipped with computational resources serve as the mobile computing layer. The proposed model ensures continuous connectivity by extending the dwell time between the service requester and the vehicular computational resource. The model ensures that the relative speed between the service vehicle (representing computational resources) and the service requester remains constant while within the communication range. We proposed a probabilistic model for the end-to-end serving time of the proposed computing paradigm. Then, we computed the dwell time between the SV and the served vehicle based on real data published by Didi Chuxing GAIA Initiative for Chengdu city, China. Utilizing a simulated model, we illustrated the additional penalty incurred by the road side unit (RSU) handovers.

The transition of the automotive sector to electric vehicles (EVs) necessitates research on charging demand forecasting for optimal station placement and capacity planning. In the literature, extensive studies have been conducted on model-based and probabilistic EV charging demand forecasting schemes. The studies provide a solid research foundation but result in complicated models with limited scalability. Meanwhile, emerging machine learning techniques bring promising prospects, yet exhibit suboptimal performance with insufficient data. Additionally, existing studies often overlook several critical areas such as overcoming data scarcity, security and privacy concerns, managing the inherent stochasticity of demand data, selecting forecasting methods for a specific feature, and developing standardized performance metrics. Considering the impact of the research topic, EV charging demand forecasting demands careful study. In this paper, we present a comprehensive survey of EV charging demand forecasting, focusing on both probabilistic and learning algorithms. First, we introduce the general procedure of EV charging demand forecasting, encompassing data sources, data pre-processing, and the key EV features. We then provide a taxonomy of existing EV charging demand forecasting techniques, followed by a critical analysis and comparative study of state-of-the-art research. Finally, we discuss open issues, which offer useful insights and future direction for various stakeholders.

This paper presents a probabilistic model for various machine learning (ML) applications. While deep learning (DL) has produced state-of-the-art results in many domains, DL models are complex and over-parameterized, which leads to high uncertainty about what the model has learned, as well as its decision process. Further, DL models are not probabilistic, making reasoning about their output challenging. In contrast, the proposed model, referred to as Yet Another Discriminate Analysis(YADA), is less complex than other methods, is based on a mathematically rigorous foundation, and can be utilized for a wide variety of ML tasks including classification, explainability, and uncertainty quantification. YADA is thus competitive in most cases with many state-of-the-art DL models. Ideally, a probabilistic model would represent the full joint probability distribution of its features, but doing so is often computationally expensive and intractable. Hence, many probabilistic models assume that the features are either normally distributed, mutually independent, or both, which can severely limit their performance. YADA is an intermediate model that (1) captures the marginal distributions of each variable and the pairwise correlations between variables and (2) explicitly maps features to the space of multivariate Gaussian variables. Numerous mathematical properties of the YADA model can be derived, thereby improving the theoretic underpinnings of ML. Validation of the model can be statistically verified on new or held-out data using native properties of YADA. However, there are some engineering and practical challenges that we enumerate to make YADA more useful.

In recent decades, Estimation of Distribution Algorithms (EDAs) have gained much popularity in the evolutionary computation community for solving optimization problems. Characterized by the use of probabilistic models to represent the solutions and the interactions between the variables of the problem, EDAs can be applied to either discrete, continuous or mixed domain problems. Due to this robustness, these algorithms have been used to solve a diverse set of real-world and academic optimization problems. However, a straightforward application is only limited to a few cases, and for the general case, an efficient application requires intuition from the problem as well as notable understanding in probabilistic modeling. In this paper, we provide a roadmap for solving optimization problems via EDAs. It is not the aim of the paper to provide a thorough review of EDAs, but to present a guide for those practitioners interested in using the potential of EDAs when solving optimization problems. In order to present a roadmap which is as useful as possible, we address the key aspects involved in the design and application of EDAs, in a sequence of stages: (1) the choice of the codification, (2) the choice of the probability model, (3) strategies to incorporate knowledge about the problem to the model, and (4) balancing the diversification-intensification behavior of the EDA. At each stage, first, the contents are presented together with common practices and advice to follow. Then, an illustration is given with an example which shows different alternatives. In addition to the roadmap, the paper presents current open challenges when developing EDAs, and revises paths for future research advances in the context of EDAs. [ABSTRACT FROM AUTHOR]

This paper proposes a probabilistic model that explains the phenomenon of cracking localization (CL) in RC beams with addition of steel fibers. Quantification of the CL is defined as the ratio between the total number of cracks and the number of significantly wide cracks. The model considers both the fibers and conventional reinforcement ratios, as well as the steel stress hardening and the location of the rebars in the cross-section. The fiber distribution in the concrete mix is considered random while the conventional reinforcement—as deterministic. A cumulative function of the total steel distribution, and a binomial probability function are proposed for a newly defined variable that represents the distribution of the fibers effectiveness along the beam. The model was validated with available data from flexural experiments showing good agreement of the model's prediction with the reported results. The model shows that the cracking localization level in beams is more pronounced in beams with low reinforcement ratios and relatively large fibers content and enables its quantification. [ABSTRACT FROM AUTHOR]

The California bearing ratio (CBR) value of subgrade is the most used parameter for dimensioning flexible and rigid pavements. The test for determining the CBR value is typically conducted under soaked conditions and is costly, labour-intensive, and time-consuming. Machine learning (ML) techniques have been recently implemented in engineering practice to predict the CBR value from the soil index properties with satisfactory results. However, they provide only deterministic predictions, which do not account for the aleatoric uncertainty linked to input variables and the epistemic uncertainty inherent in the model itself. This work addresses this limitation by introducing an ML model based on the natural gradient boosting (NGBoost) algorithm, becoming the first study to estimate the soaked CBR value from this probabilistic perspective. A database of 2130 soaked CBR tests was compiled for this study. The NGBoost model showcased robust predictive performance, establishing itself as a reliable and effective algorithm for predicting the soaked CBR value. Furthermore, it produced probabilistic CBR predictions as probability density functions, facilitating the establishment of reliable confidence intervals, representing a notable improvement compared to conventional deterministic models. Finally, the Shapley additive explanations method was implemented to investigate the interpretability of the proposed model. [ABSTRACT FROM AUTHOR]

The gust factor commonly is used in wind engineering community to convert mean wind speeds into gusty winds, which exhibit significant variability in real observations of typhoon winds. This study proposes a probabilistic gust factor model that accounts for uncertainties of wind speed statistics. The statistical characteristics of wind speed from nine typhoons, including the mean, standard deviation, skewness, kurtosis, power spectral density (PSD) parameter, peak factor, and gust factor, were examined. The effects of nonstationary characteristics in terms of time-varying mean wind speed and non-Gaussian attributes of fluctuating winds on the gust factor are discussed. These wind speed statistics were incorporated into a non-Gaussian moment-based translation model to perform the Monte Carlo simulation of peak factor and gust factor. The simulation results for different gust durations were juxtaposed with observations to substantiate the accuracy of the probabilistic model. Subsequently, a standardization framework for estimating site-specific probabilistic gust factor curves was developed. This approach was applied to determine the gust typhoon wind speed hazard curve at a real bridge site with flat open terrain. The present model enables the consideration of gust factor dispersion to achieve a probabilistic gust wind hazard curve and facilitate the development of performance-based wind engineering. [ABSTRACT FROM AUTHOR]

The geological modeling commonly performed on mineral deposits is of the deterministic type, but it has been demonstrated that probabilistic modeling is more adequate to know or quantitatively measure the associated uncertainty. This study is aim to quantify the uncertainty of weathering crustal horizons for a sector of the San Felipe nickel deposit, based on its probabilistic modeling. The descriptive geological model of the lateritic deposit was taken as a basis, characterized by the lithological zoning at depth. The probabilistic Sequential Indicator Simulation with Locally variable Proportions was used, through which 50 equiprobable scenarios of the geological units are synthetically reproduced, conditioned to 3,884 composites from the network of 100 m x 100 m drillholes. As a result, the 3D model of the lithological horizons on block support is presented for the study sector, as well as the entropy maps that quantify the uncertainty associated with the model of the modeled geological units. As conclusion, we can say that, due to the variability present in this type of lateritic deposit, the availability of geological information is still insufficient to reproduce the model of weathering crust horizons with an adequate level of certainty. [ABSTRACT FROM AUTHOR]

Background and Objectives: The detection of community in networks is an important tool for revealing hidden data in network analysis. One of the signs that the community exists in the network is the neighborhood density between nodes. Also, the existence of a concept called a motif indicates that a community with a high edge density has a correlation between nodes that goes beyond their close neighbors. Motifs are repetitive edge patterns that are frequently seen in the network. Methods: By estimating the triangular motif in the network, our proposed probabilistic motif-based community detection model (PMCD) helps to find the communities in the network. The idea of the proposed model is network analysis based on structural density between nodes and detecting communities by estimating motifs using probabilistic methods. Results: The suggested model's output is the strength of each node's affiliation to the communities and detecting overlaps in communities. To evaluate the performance and accuracy of the proposed method, experiments are done on real-world and synthetic networks. The findings show that, compared to other algorithms, the proposed method is acting more accurately and densely in detecting communities. Conclusion: The advantage of PMCD in using the probabilistic generative model is speeding up the computation of the hidden parameters and establishing the community based on the likelihood of triangular motifs. In fact, the proposed method proves there is a probabilistic correlation between the observation of two node pairs in different communities and the increased existence of motif structure in the network. [ABSTRACT FROM AUTHOR]

Urban Air Mobility (UAM) is an emerging form of transportation that is expected to introduce novel flight networks into already busy and complex airspace surrounding major cities and metropolitan regions. This paper studies the dynamics of urban airspace use by conventional aircraft over the Sao Paulo metropolitan region in order to identify and predict which airspace volumes are least constrained and best accessible for future UAM flights. Using historical flight tracking data, clustering analysis is first performed to identify departure and arrival trajectory patterns flown by conventional traffic at the two major airports – Sao Paulo/Guarulhos International airport and Sao Paulo/Congonhas airport. We then create a probabilistic model of the spatiotemporal distribution of air traffic under known meteorological conditions, which enables the prediction of active procedures, their spatial confidence regions and the resulting airspace availability for UAM in response to dynamic operational factors. The data-based approach allowed for a high-fidelity characterization of the Sao Paulo urban airspace use patterns as well as for accurate predictions of the available airspace for UAM, bringing novel insights and capabilities in support of dynamic and efficient urban airspace management.

Objectives Running a business effectively requires ensuring continuity and regularity of the activity. For this purpose, it is important to effectively and rationally manage inventories and maintain them at various levels of optimality. Material and methods In the paper a design of the probabilistic model of reserves is discussed. The objective was an elaboration of the optimal strategy of materials reserves management in the series manufacturing. Results The model has been verified at an attainable scale using the numerical data concerning different assortment of materials in the serial production of furniture. The proposed probabilistic model makes it possible to elaborate the optimal strategy of type R, Z in the management of reserves of based materials of the manufacturing company with large-lot production. The strategy is based on minimal expenses connected with a supply of materials. Conclusions For each company to provide good quality services, it is necessary for individual, cooperating enterprises to function efficiently. The so-called "supply chain" is a specific sequence of activities enabling the satisfaction of market demand for a given product. The supply chain consists of companies and plants that are involved in supplying raw materials, processing them into semi-finished products and, ultimately, creating a finished product. The simplest supply chain consists of a company, suppliers and customers. However, it is good to know that many more companies are involved in most production processes, including transport, logistics, finance and IT.

Model-based hyperspectral image (HSI) denoising methods have attracted continuous attention in the past decades, due to their effectiveness and interpretability. In this work, we aim at advancing model-based HSI denoising, through sophisticated investigation for both the fidelity and regularization terms, or correspondingly noise and prior, by virtue of several recently developed techniques. Specifically, we formulate a novel unified probabilistic model for the HSI denoising task, within which the noise is assumed as pixel-wise non-independent and identically distributed (non-i.i.d) Gaussian predicted by a pre-trained neural network, and the prior for the HSI image is designed by incorporating the deep image prior (DIP) with total variation (TV) and spatio-spectral TV. To solve the resulted maximum a posteriori (MAP) estimation problem, we design a Monte Carlo Expectation–Maximization (MCEM) algorithm, in which the stochastic gradient Langevin dynamics (SGLD) method is used for computing the E-step, and the alternative direction method of multipliers (ADMM) is adopted for solving the optimization in the M-step. Experiments on both synthetic and real noisy HSI datasets have been conducted to verify the effectiveness of the proposed method.