Variable Time Research Articles

Quality and applicability of cadaveric donor eyes for molecular biology research: An Indian experience.

Human ocular tissue banking plays an important part in the advancement of translational research for identifying the molecular processes involved in disease etiology and pathogenesis. Timely obtaining a good-quality ocular tissue from a cadaveric donor is exceedingly difficult, especially in remote areas, with a variable transportation time (within 12-24 h), raising concerns about RNA quality and its subsequent applications. Therefore, we assessed the utility of retinal tissues from cadaver donor and enucleated eyes based on the RNA quality and gene expression by real-time polymerase chain reaction (PCR). Prospective study. Retina tissues were separated from the donor/enucleated eyes received in the eye bank within 24 h of death (n = 15) and within an hour from OR (n = 3), respectively, and stored immediately at -80 degree. RNA was isolated using trizol, and the quantity and quality were assessed using Qubit and agarose gel electrophoresis, respectively. QPCR was performed for measuring the expression of different retinal-specific genes. The cellular viability of the retina was assessed by establishing explant primary cell cultures. The data were calculated as an average of normalised Ct values ± standard error of the mean. RNA obtained from cadaveric tissues despite being partially degraded showed a uniform strong gene expression of several retinal-specific genes such as PAX6, RHO, TUBB3, CRX , and ALDH1L1 . The primary cultures established from cadaveric tissues showed viable cells. The cadaver donor tissues collected within 24 hours of death can be effectively utilized for gene expression profiling.

Using Machine Learning Techniques to Forecast Mehram Company's Sales: A Case Study

Sales forecasting, situated at the intersection of art and science, is critical for inspiring managers toward achieving profitable outcomes. Its precision sustains production levels and capital and plays a pivotal role in the company's and its leaders' overall success and career progression. In the context of Mahram Food Industries, the challenge arises from diverse investor perspectives and the impactful nature of numerous variables. To address this, a new sales forecasting algorithm has been introduced to enhance accuracy. The aim is to predict future sales through a comprehensive approach, leveraging technical analysis, time series modeling, machine learning, neural networks, and random forest techniques. The research methodology integrates various advanced techniques to improve sales forecasting for Mahram Food Industries. Technical analysis, time series modeling, machine learning, neural networks, and random forest methods are combined to create a robust framework. The focus is on predicting sales for a future period within the artificial intelligence-based machine learning domain. The study employs metrics such as Mean Absolute Deviation (MAD), MAD Percentage (MADP), and Mean Squared Error (MSE) to evaluate and compare the performance of the proposed neural network against traditional methods like multiple variable regression and time series modeling. The study's results highlight the superior performance of the neural network in sales forecasting for Mahram Food Industries. The Mean Absolute Deviation (MAD) for the neural network is 28.33, outperforming multiple variable regression (28.54) and time series modeling (29.45). Additionally, the neural network demonstrates a better MAD Percentage (MADP) with a value of 10.2%, surpassing the values associated with multiple variable regression (10.35%) and time series modeling (10.30%). The Mean Squared Error (MSE) further confirms the neural network's superiority with a value of 6452 compared to 6472 and 7865 for multiple variable regression and time series modeling, respectively. In conclusion, the study showcases the effectiveness of integrating advanced techniques, particularly the neural network, in enhancing the accuracy of sales forecasting for Mahram Food Industries. The comprehensive approach, combining technical analysis, time series modeling, machine learning, neural networks, and random forest, is a valuable strategy for predicting future sales. The superior performance of the neural network, as evidenced by lower MAD, MADP, and MSE values, suggests its potential for guiding informed decision-making in goal setting, hiring, budgeting, and other critical aspects of business management.

Solid state NMR study of the bioenhancer piperine, supported by GIPAW DFT calculations

Piperine, an alkaloid found in black pepper fruits, has the properties of promoting the absorption of other substances (e.g. curcumin), therefore it is used in solid forms of dietary supplements as an additive increasing bioavailability. The aim of the study was to analyze piperine in a solid state and then to use it in the analysis of solid dietary supplements with the addition of piperine as an absorption promoter. The 13C CP MAS NMR spectra were recorded using variable contact time and dipolar dephasing experiment. The calculation of theoretical chemical shift values for three polymorphic forms of piperine allowed us to create a database of NMR parameters enabling the identification of polymorphic forms based on the analysis of the 13C CP MAS NMR spectrum. Additionally, the analysis of cross-polarization kinetics parameters was performed. Then, the 13C CP MAS NMR technique was used to confirm the authenticity and determine the presence of curcumin in dietary supplements containing curcumin with the addition of piperine. The presence of piperine could be confirmed even when the mass content of piperine was 70 times lower than that of curcumin. This method can be used to test the quality of dietary supplements containing the addition of piperine as an absorption promoter.

Hormonal Dysfunction in Paediatric Patients Admitted to Rehabilitation for Severe Traumatic Brain Injury: Analysis of the Associations with Rehabilitation Outcomes.

Traumatic brain injury is often accompanied by defects in hormone levels, caused by either peripheral gland dysfunctions or by an insufficient central stimulation of hormone production. The epidemiology of endocrinological defects after traumatic brain injury is quite well described, but the consequences of hormone defects are largely unknown, especially in paediatric patients undergoing neurological rehabilitation. Only one previous study reported on a cohort of 20 children with traumatic brain injury and found a low incidence of hormone defects and a correlation between some hormone levels and neurological recovery. In this study, we performed a retrospective chart review on patients affected by severe subacute traumatic brain injury. Their levels of cortisol, ACTH, IGF-1, TSH, free T4, free T3, and prolactin were collected and compared with reference ranges; we then used regression models to highlight any correlation among them and with clinical variables; last, we probed with regression models whether hormone levels could have any correlation with clinical and rehabilitation outcomes. We found eligible data from the records of 52 paediatric patients with markedly severe traumatic brain injury, as shown by an average GCS of 4.7; their age was 10.3 years, on average. The key results of our study are that 32% patients had low IGF-1 levels and in multiple regression models, IGF-1 levels were correlated with neurological recovery, indicating a possible role as a biomarker. Moreover, 69% of patients had high prolactin levels, possibly due to physical pain and high stress levels. This study is limited by the variable timing of the IGF-1 sampling, between 1 and 2 months after injury. Further studies are required to confirm our exploratory findings.

The control method for ship tracking when navigating through narrow and curved sections

A track control method is proposed to solve the problem of manual steering difficulties in narrow and curved sections. The performance of a ship heading controller is affected by various factors, such as delay of the ship steering gear, external environmental disturbances, and variations in ship's speed. Firstly, a Smith optimal PID heading controller is designed. Then, a speed control method based on dichotomous method is proposed to accurately control ship speed and improve the rudder efficiency. Finally, a track control method based on model prediction and feedback compensation is proposed, considering the coupling relationship between course control and speed control. The Lantang Channel is selected as the experimental water area, and comparative simulation experiments are carried out under static water and environmental disturbance. The pre-planned route information is taken as input, and the target course at the current moment is calculated by the guidance method. The target course is used as the input of the course controller to calculate the rudder angle to control the ship's motion, and the new ship position is compared with the planned route information. The above steps are repeated to track the planned route. In order to make the ship pass the curve smoothly, the target speed of the control point is preset according to the sailing convention of decelerating first and accelerating later, and the variable speed time point and variable telegraph are solved by the dichotomy. The results are as follows: (1) The maximum track deviation and speed deviation of the Smith-OP PID are 2 m and 0.21 kn, respectively, without considering environmental interference. The maximum track deviation of the traditional PID is 8 m. (2) The maximum track deviation and speed deviation of the Smith-OP PID are 25 m and 0.27 kn, respectively, under the set environmental interference. The maximum track deviation of the traditional PID is 49 m. (3) The track control method is consistent with the practical practices of "deceleration before acceleration" and "hanging high and taking low". In summary, the track control method can be well applied to narrow and curved sections, and provide a theoretical basis for subsequent research on autonomous navigation and intelligent ships.

Boundedness of Vector Linéard Equation with Multiple Variable Delays

In this article, we consider a system of ordinary differential equations (ODEs) of second order with two variable time delays. We obtain new conditions for uniform ultimate bounded (UUB) solutions of the considered system. The technique of the proof is based on the Lyapunov–Krasovskii functional (LKF) method using a new LKF. The main result of this article extends and improves a recent result for ODEs of second order with a constant delay to a more general system of ODEs of second order with two variable time delays. In this particular case, we also give a numerical example to verify the application of the main result of this article.

Viscoelastic modeling via fractional calculus of the cold bending of laminated glass

A viscoelastic description via fractional calculus is used to theoretically determine the time-varying stress state in single-curvature cold-bent laminated glass. This approach is proven effective when the relaxation function of the polymeric interlayer can be approximated by branches of power laws, as in most commercial materials. Solutions are obtained numerically by approximating the fractional time derivatives with the L1 formula. This conveniently allows to use a variable time step for a phenomenon characterized by two time-scales, corresponding to the loading process and the long-term relaxation. A parametric analysis shows the effects of polymer type, interlayer thickness, deformation history and operating temperature. A comparison is made with the results from the quasi-elastic approximation, which neglects the memory effect of viscoelasticity, showing that, since the interlayer strain is kept constant in the long term, it provides accurate results in term of peak and asymptotic stresses in glass, although the early stages of the deformation history give rise to differences when the structure is far from the layered or monolithic limits. This study can be useful for the optimization of the cold-bending process.

Numerical model of variable valve timing distribution for a supercharged diesel engine

Recently, there's been a strong drive to improve performance of diesel engines while reducing their greenhouse gases emissions. Techniques like exhaust gas recirculation, turbocharging, and variable valve timing have become widespread. The last technique fine-tunes valve operation based on engine speed, which optimize efficiency and power output while saving fuel. This study zeroes in on a specific 4-cylinder, 4-stroke diesel engine of 1.56-liter, GT-Power software is employed to examine a supercharged version and implementing diverse valve lift techniques. The findings are revealing a substantial 30% increase in power output. At 1000 rpm, power rises from 15.1 kW for the standard engine to 19.72 kW for the modified version. For higher engine speeds, the improvements become even more pronounced, reaching a 66% boost compared to the standard configuration. Furthermore, the newly configured engine showcases an impressive 13% decrease in fuel-specific consumption at elevated engine speeds, contributing to enhanced technical performance and fuel efficiency. The numerical model developed in this study holds the potential to aid in the design of novel diesel engines equipped with variable valve timing systems. To lend further support to these findings, experimental validation is recommended.

Automatic smooth map generation of internal combustion engines via local-global model based calibration technique

The addition of new sensors and actuators to the engine, to reduce fuel consumption and emissions besides improving the engine operation, complicates the control commands stored in the engine control unit (ECU). Substitution of mechanical actuators with electronic ones increases the engine’s degrees of freedom and the number of control parameters, which results in the increased engine calibration time and cost. The aim of this paper is to take advantage of optimization techniques to achieve optimal values of control parameters in a fast and automated way. In this regard, it requires replacing the real engine with the virtual model and implementing the model-based calibration by coupling the virtual engine model with optimization algorithms. In this study, deep neural network (DNN) modeling and genetic algorithm (GA, NSGA-II) optimization are used for model-based calibration. The effect of all input control parameters, including ignition angle, continuously variable valve timing, etc., on all output control parameters including, brake-specific fuel consumption, emissions level, knock limit, combustion stability, etc., are investigated simultaneously by a valid global model, which is a remarkable achievement in the model-based calibration. Dynamic lag of some actuators delays the execution of control commands sent from ECU. To avoid abrupt variations in the actuators values, smoothness of the engine maps is considered in the calibration process. To reduce fuel consumption, decrease emission levels and attain smooth maps, the calibration of control parameters is performed by local-multi-objective optimization and global-single-objective optimization. Local-global model-based calibration presented in this study reduces 3.7% of the brake-specific fuel consumption and 7%–10% of emissions level at breakpoints of the engine map compared to manual calibration. In addition, the calibration time and costs while producing better engine performance can be reduced by automating the calibration process. Finally, calibrated maps are stored as a lookup table (LUT) in ECU. Generating an optimal lookup table involves the pre-calculation of several points that cover the calculation domain and allow the interpolation for other points. Selecting the optimal points for exact calculation is of great importance in the size and accuracy of LUT. In this study, an optimization tool is also presented to generate accurate and efficient LUT.

Anticoagulant Rodenticide Toxicity in Terrestrial Raptors: Tools to Estimate the Impact on Populations in North America and Globally.

Anticoagulant rodenticides (ARs) have caused widespread contamination and poisoning of predators and scavengers. The diagnosis of toxicity proceeds from evidence of hemorrhage, and subsequent detection of residues in liver. Many factors confound the assessment of AR poisoning, particularly exposure dose, timing and frequency of exposure, and individual and taxon-specific variables. There is a need, therefore, for better AR toxicity criteria. To respond, we compiled a database of second-generation anticoagulant rodenticide (SGAR) residues in liver and postmortem evaluations of 951 terrestrial raptor carcasses from Canada and the United States, 1989 to 2021. We developed mixed-effects logistic regression models to produce specific probability curves of the toxicity of ∑SGARs at the taxonomic level of the family, and separately for three SGARs registered in North America, brodifacoum, bromadiolone, and difethialone. The ∑SGAR threshold concentrations for diagnosis of coagulopathy at 0.20 probability of risk were highest for strigid owls (15 ng g-1 ) lower and relatively similar for accipitrid hawks and eagles (8.2 ng g-1 ) and falcons (7.9 ng g-1 ), and much lower for tytonid barn owls (0.32 ng g-1 ). These values are lower than those we found previously, due to compilation and use of a larger database with a mix of species and source locations, and also to refinements in the statistical methods. Our presentation of results on the family taxonomic level should aid in the global applicability of the numbers. We also collated a subset of 440 single-compound exposure events and determined the probability of SGAR-poisoning symptoms as a function of SGAR concentration, which we then used to estimate relative SGAR toxicity and toxic equivalence factors: difethialone, 1, brodifacoum, 0.8, and bromadiolone, 0.5. Environ Toxicol Chem 2024;00:1-11. © 2024 His Majesty the King in Right of Canada and The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC Reproduced with the permission of the Minister of Environment and Climate Change Canada.

New method for collecting XPS and other spectra: A thought (Gedanken) experiment

We suggest a new approach for collecting data, with accompanying statistical analysis, that could be useful for assessing sample damage in X-ray Photoelectron Spectroscopy (XPS) and for target detection in XPS imagining. It is based on the concept of the ‘constant signal, variable time’ (CSVT) spectrum that is in many ways the opposite of the traditional approach of collecting ‘variable signal, constant time’ spectra. In many cases, CSVT spectra should allow better visual comparisons of spectra and also for more rigorous statistical analyses based on paired t-tests and sums-of-squares of errors. We discuss the possibility of taking both survey and narrow CSVT spectra and provide a statistical discussion of how the resulting data might be treated. As we have no ability to collect CSVT spectra with our XPS instrument, this paper is a thought (Gedanken) experiment. The ideas presented herein may also be transferable to other spectroscopies.

A comprehensive assessment and comparison of the impacts of storage parameters on solute transport in streams using a novel framework

Researchers must determine the storage parameter values to understand the storage processes and compare them across streams. These parameters are highly variable in time and space, making them challenging to measure. Individual studies of a single stream cannot characterize the interrelations among the processes involved in solute transport, hydraulic variables, and physical stream characteristics. Despite decades of research, a comprehensive framework is lacking. This study uses a simple and effective Monte Carlo approach to propose a general framework within which most solute transport models can undergo sensitivity analysis. This framework provides a comprehensive depiction of model performance under various scenarios, such as flow properties, stream geomorphology, and solute status. The relative and absolute sensitivities of three solute transport models, namely, the modified advection–dispersion equation (MADE), transient storage model (TSM), and variable residence time (VART), were evaluated and compared within the proposed framework. The main findings can be summarized as follows. In most conditions, storage models are more sensitive to the ratio of the storage zone cross-section to the stream cross-section than other parameters. Increasing the exchange rate between the stream and storage zones reduces the sensitivity to its indicative parameter and, consequently, its impact on the concentration time series. MADE's capability to identify modification parameters is higher for streams where dispersion dominates over advection, and their estimates have lower uncertainty. On the other hand, as advection becomes more dominant over dispersion in TSM and VART, estimated storage parameters have lower uncertainty.“

Short-term trajectory prediction for individual metro passengers based on multi-level periodicity mining from semantic trajectory

Short-term trajectory prediction for individual metro passengers is an essential component of many location-based applications, such as bottleneck identification, anomalous trajectory detection, and emergency evacuation. Accurate and interpretable trajectory prediction in short-term remains challenging since the trajectory of individual metro passengers is influenced by time-varying and periodic mobility patterns. Moreover, it is still difficult for deep learning-based prediction approaches to intuitively explain the individual periodic mobility patterns captured by the black-box model. In this study, a long short-term memory method based on multi-level periodicity mining (LSTM-MP) from semantic trajectory is developed to achieve accurate and interpretable short-term trajectory prediction for individual metro passengers. The newly defined semantic trajectories of metro passengers that combine the semantic information about the geometric and operational characteristics of the metro with the raw trajectory data are constructed to help obtain fine-grained periodic mobility patterns of individual passengers. Multi-level periodicity which includes long-range, weekly-interval, and daily-interval periodicities, is calculated with variable time granularity and its polymerization from the constructed semantic trajectories. Finally, the characteristics of multi-level periodicity are considered by developing a principal trajectory attention mechanism to select the most relevant trajectories for prediction in the LSTM-MP framework. The prediction performance of LSTM-MP is compared with 2-order Markov chain and Naïve-LSTM algorithms using the large real-world anonymous passenger WiFi probe dataset. Experimental results show that LSTM-MP can predict when and where an individual will be in the future time series with an average accuracy of 82.3%. This demonstrates the effectiveness of LSTM-MP in the short-term trajectory prediction for individual metro passengers. Potential applications of the short-term trajectory prediction for individual metro passengers are also identified.

Isotopic tracing of leachate percolation from municipal solid waste dump sites to groundwater in diverse climatic zones of India.

Groundwater resources in tropical regions are largely dependent on recharge by rainwater infiltration through soil layers with variable time. However, the rainwater infiltration through soil is a serious concern in urban tropics where it interacts with landfills at the dumpsites, potentially contaminating adjoining groundwater. In this study, the stable isotopic compositions of oxygen and hydrogen (δ18O and δ2H, respectively) in groundwater and leachates, adjoining municipal dumpsites in urban tropics (Bangalore, Kolkata and Durgapur located in diverse rainfall zonation of India), were analyzed to investigate their recharge sources and trace the possible mixing of leachate contaminants under three diverse climatology. The measured values of δ18O and δ2H suggested that the groundwater in these sites reflects higher recharge by rainwater. However, the d-excess values indicated secondary effects suggesting the groundwater has experienced significant modifications. The end member analysis using δ18O-d-excess relation pinpointed an additional leachate contribution from adjoining dumpsites. The critical fraction of leachate infiltration to groundwater quantified using two component mixing model ranged between (i) 1 and 33% in Bangalore, (ii) 5 and 13% in Kolkata and (iii) 18 and 76% in Durgapur, with its variability dependent on seasonality and aquifer connectivity. This information is crucial for groundwater management to secure water quality and to quantify potential hydrological contaminants particularly in drier seasons and drier regions, when and where the demand for groundwater is high, respectively.

Participatory modeling for high complexity, multi‐system issues: challenges and recommendations for balancing qualitative understanding and quantitative questions

AbstractCommunity stakeholder participation can be incredibly valuable for the qualitative model development process. However, modelers often encounter challenges for participatory modeling projects focusing on high‐complexity, synergistic interactions between multiple issues, systems, and granularity. The diverse stakeholder perspectives and volumes of information necessary for developing such models can yield qualitative models that are difficult to translate into quantitative simulation or clear insight for informed decision‐making. There are few recommended best practices for developing high‐complexity, participatory models. We use an ongoing project as a case study to highlight three practical challenges for tackling high‐complexity, multi‐system issues with system dynamics tools. These challenges include balanced and respectful stakeholder engagement, defining boundaries and levels of variable aggregation, and timing and processes for qualitative/quantitative model integration. Our five recommendations to address these challenges serve as a foundation for further research on methods for developing translatable qualitative multi‐system models for informing actions for systemic change. © 2024 System Dynamics Society.

Application of a Variable Weight Time Function Combined Model in Surface Subsidence Prediction in Goaf Area: A Case Study in China

To attain precise forecasts of surface displacements and deformations in goaf areas (a void or cavity that remains underground after the extraction of mineral resources) following coal extraction, this study based on the limitations of individual time function models, conducted a thorough analysis of how the parameters of the model impact subsidence curves. Parameter estimation was conducted using the trust-region reflective algorithm (TRF), and the time function models were identified. Then we utilized a combined model approach and introduced the sliding window mechanism to assign variable weights to the model. Based on this, the combined model was used for prediction, followed by the application of this composite prediction to engineering scenarios for the dynamic forecasting of surface movements and deformations. The results indicated that, in comparison with DE, GA, PSO algorithms, the TRF exhibited superior stability and convergence. The parameter models obtained using this method demonstrated a higher level of predictive accuracy. Moreover, the predictive precision of the variable-weight time function combined model surpassed that of corresponding individual time function models. When employing six different variable-weight combination prediction models for point C22, the Weibull-MMF model demonstrated the most favorable fitting performance, featuring a root mean square error (RMSE) of 32.98 mm, a mean absolute error (MAE) of 25.66 mm, a mean absolute percentage error (MAPE) of 7.67%; the correlation coefficient R2 reached 0.99937. These metrics consistently outperformed their respective individual time function models. Additionally, in the validation process of the combined model at point C16, the residuals were notably smaller than those of individual models. This reaffirmed the accuracy and reliability of the proposed variable-weight combined model. Given that the variable-weight combination model was an evolution from individual time function models, its applicability extends to a broader range, offering valuable guidance for the dynamic prediction of surface movement and deformation in mining areas.

Analyzing the Factors Influencing Time Delays in Korean Railroad Accidents

Railroads play a pivotal role in the Korean national economy, necessitating a thorough understanding of factors influencing accidents for effective mitigation strategies. Unlike prior research focused on accident frequency and severity, this study delves into the often-overlooked aspect of time delays resulting from railroad accidents. Analyzing 15 years of nationwide data (2008–2022), encompassing 3244 human-related and 3350 technical events, this research identifies key factors influencing delay likelihood and duration. Factors considered include event type, season, train type, location, operator size, person type involved, facility type, and causes. Despite an overall decrease in events, variable delay times highlight the need to comprehend specific contributing factors. To address excess zeros, the study employs a two-stage model and a zero-inflated negative binomial (ZINB) model, alongside artificial neural networks (ANNs) for non-linear pattern recognition. Human-related delays are influenced by event types, seasons, and passenger categories, exhibit nuanced impacts. Technical-related delays are influenced by incident types and facility involvement. Regarding model performance, the ANN models outperform regression-based models consistently in all cases. This study emphasizes the importance of considering both human and technical factors in predicting and understanding railroad accident delays, offering valuable insights for formulating strategies to mitigate service disruptions associated with these incidents.

Physics-informed neural nets for control of dynamical systems

Physics-informed neural networks (PINNs) incorporate established physical principles into the training of deep neural networks, ensuring that they adhere to the underlying physics of the process while reducing the need for labeled data, since the desired output is not a prerequisite for physics-informed training. For modeling systems described by Ordinary Differential Equations (ODEs), traditional PINNs typically take continuous time as an input variable and produce the solution to the corresponding ODE. However, in their original form, PINNs neither accommodate control inputs nor do they effectively simulate variable long-range intervals without experiencing a significant decline in prediction accuracy. In this context, this work introduces a novel framework known as “Physics-Informed Neural Nets for Control” (PINC). PINC presents an innovative PINN-based architecture tailored to control problems, capable of simulating longer-range time horizons that are not predetermined during training. This increased flexibility sets it apart from traditional PINNs. The variable simulation time is achieved by adding inputs to the PINC network that convey the initial condition and the control signal for a particular time interval. Simulating variable long-range intervals involves running the PINC net across a sequence of shorter intervals. In this autoregressive process, the network predictions are linked in a self-feedback mode, with the initial state (input) of the next interval set to the last predicted state (network output) of the previous interval. We showcase the effectiveness of our proposal in identifying and controlling three nonlinear dynamic systems: the Van der Pol oscillator, the four-tank system, and an electric submersible pump. Crucially, these experiments demonstrate that learning the dynamics of these systems can be achieved without relying on any sample collected from the actual process, and it offers faster inference speed compared to numerical simulations.

Research on decision making of intelligent vehicle based on composite priority experience replay

To address the problems of underutilization of samples and unstable training for intelligent vehicle training in the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm, a TD3 algorithm based on the Composite Prioritized Experience Replay (CPR-TD3) mechanism is proposed. It considers experience immediate reward value and Temporal Difference error (TD-error) based and respectively to construct priorities to rank the samples. Subsequently composite average ranking of the samples to recalculate the priorities for sampling, uses the collected samples to train the target network. Then introduces the minimum lane change distance and the variable headway time distance to improve the reward function. Finally, the improved algorithm is proved to be effective by comparing it with the traditional TD3 on the highway scenario, and the CPR-TD3 algorithm improves the training efficiency of intelligent vehicles.

An end-to-end machine learning approach with explanation for time series with varying lengths

AbstractAn accurate prediction of complex product quality parameters from process time series by an end-to-end learning approach remains a significant challenge in machine learning. A special difficulty is the application of industrial batch process data because many batch processes generate variable length time series. In the industrial application of such methods, explainability is often desired. In this study, a 1D convolutional neural network (CNN) algorithm with a masking layer is proposed to solve the problem for time series of variable length. In addition, a novel combination of 1D CNN and class activation mapping (CAM) technique is part of this study to better understand the model results and highlight some regions of interest in the time series. As a comparative state-of-the-art unsupervised machine learning method, the One-Nearest Neighbours (1NN) algorithm combined with dynamic time warping (DTW) was used. Both methods are investigated as end-to-end learning methods with balanced and unbalanced class distributions and with scaled and unscaled input data, respectively. The FastDTW and DTAIDistance algorithms were investigated for the DTW calculation. The data set is made up of sensor signals that was collected during the production of plastic parts. The objective was to predict a quality parameter of plastic parts during production. For this research, the quality parameter will be a difficult or only destructively measurable parameter and both methods will be investigated for their applicability to this prediction task. The application of the proposed approach to an industrial facility for producing plastic products shows a prediction accuracy of 83.7%. It can improve the reverence method by approximately 1.4%. In addition to the slight increase in accuracy, the CNN training time was significantly reduced compared to the DTW calculation.