Explore the vast range of titles available.

A significant number of sample points are often required for surrogate-based optimization when utilizing process simulations to cover the entire system space. This necessity is particularly pronounced in complex simulations or high-dimensional physical experiments, where a large number of sample points is essential. In this study, we have developed an adaptive Latin hypercube sampling (LHS) method that generates additional sample points from areas with the highest output deviations to optimize the required number of samples. The surrogate model used for the optimization problem is artificial neural networks (ANNs). The standard for measuring solution accuracy is the percent error of the optimal solution. The outcomes of the proposed algorithm were compared to those of random sampling for validation. As case studies, we chose three different chemical processes to illustrate problems of varying complexity and numbers of variables. The findings indicate that for all case studies, the proposed LHS optimization algorithm required fewer sample points than random sampling to achieve optimal solutions of similar quality. To extend the application of this methodology, we recommend further applying it to fields beyond chemical engineering and higher-dimensional problems.

A novel, yet practically feasible, robust optimization approach is proposed in this study for engineering structures involving hybrid uncertainties. Both stochastic and interval uncertain system parameters are incorporated within a single analysis-design computational scheme. The generalized beta distribution is adopted to model the bounded stochastic system uncertainties, which offers the benefit of evaluating the performance of objective function and constraints of the robust optimization. A multi-layered refining Latin hypercube sampling–based Monte Carlo simulation approach is proposed to assess the robustness of the objective function. Furthermore, a new concept, namely, the interval angular vector, is presented to evaluate the robust feasibility of the constraints of the optimization problem. In order to systematically solve the robust optimization problem, a new genetic algorithm is presented in this study which utilizes the order preference by similarity to ideal solution technique so the feasible design vectors can be sorted according to their distances to the negative ideal solution. The effectiveness and applicability of the proposed computational approach are demonstrated by one numeral example and two realistic complex engineering structures including the bucket linkage mechanism of an excavator and the upper beam of a high-speed punching machine.

A probabilistic load flow (PLF) method using Copula and improved Latin hypercube sampling is proposed. The stochastic dependence between input random variables is considered. Copula theory is adopted to establish the probability distribution of correlated input random variables. Based on discrete data, an improved Latin hypercube sampling is proposed. The accuracy of probability distribution of correlated input random variables established by Copula theory is evaluated by adopting the power output of wind farms located at New Jersey. The performance of the proposed PLF method is investigated using IEEE 14-bus and IEEE 118-bus test systems.

Bayesian optimization (BO)-assisted screening was applied to identify improved reaction conditions toward a hundred-gram scale-up synthesis of 2,3,7,8-tetrathiaspiro[4.4]nonane (1), a key synthetic intermediate of 2,2-bis(mercaptomethyl)propane-1,3-dithiol [tetramercaptan pentaerythritol]. Starting from the initial training set (ITS) consisting of six trials sampled by random screening for BO, suitable parameters were predicted (78% conversion yield of spiro-dithiolane 1) within seven experiments. Moreover, BO-assisted screening with the ITS selected by Latin hypercube sampling (LHS) further improved the yield of 1 to 89% within the eight trials. The established conditions were confirmed to be satisfactory for a hundred grams scale-up synthesis of 1.

Sensitivity analysis of hydrological model parameters is a crucial step in the calibration process of hydrological simulation. In this paper, the improved Morris method with the double-Latin hypercube sampling is proposed for global sensitivity analysis of 10 parameters of the Xin'anjiang model. In addition, the local sensitivity is analyzed based on the rate validation of the model parameters. In general, the results show those parameters about evaporation coefficient in the deep layer (C), free water storage capacity (SM), impervious area as a percentage of total watershed area (IMP), free water storage capacity curve index (EX), groundwater outflow coefficient (KG) and subsurface runoff abatement factor (KKG) are all less than 0.01, insensitive parameters; the parameters about evaporation conversion factor (K) and square times of the storage capacity curve(B) are in the range of [0.01, 0.1], less sensitive parameters; the parameter for flow out coefficient in soil (KSS) is in the range of [0.1, 0.2], a low-sensitivity parameter; the parameter abatement coefficient of mid-soil flow (KKSS) is greater than 1, a high-sensitivity parameter; the improved Morris method better reflects the existence of interactions between parameters. This research result provides a new technical approach for the sensitivity analysis of hydrological model parameters.

Background Tuberculosis (TB), a preventable and curable disease, is claimed as the second largest number of fatalities, and there are 9,025 cases reported in the United States in 2018. Many researchers have done a lot of research and achieved remarkable results, but TB is still a severe problem for human beings. The study is a further exploration of the prevention and control of tuberculosis. Methods In the paper, we propose a new dynamic model to study the transmission dynamics of TB, and then use global differential evolution and local sequential quadratic programming (DESQP) optimization algorithm to estimate parameters of the model. Finally, we use Latin hypercube sampling (LHS) and partial rank correlation coefficients (PRCC) to analyze the influence of parameters on the basic reproduction number ( R 0 ) and the total infectious (including the diagnosed, undiagnosed and incomplete treatment infectious), respectively. Results According to the research, the basic reproduction number is computed as 2.3597 from 1984 to 2018, which means TB is also an epidemic in the US. The diagnosed rate is 0.6082, which means the undiagnosed will be diagnosed after 1.6442 years. The diagnosed will recover after an average of 1.9912 years. Moreover, some diagnosed will end the treatment after 1.7550 years for some reason. From the study, it’s shown that 2.40% of the recovered will be reactivated, and 13.88% of the newborn will be vaccinated. However, the immune system will be lost after about 19.6078 years. Conclusion Through the results of this study, we give some suggestions to help prevent and control the TB epidemic in the United States, such as prolonging the protection period of the vaccine by developing new and more effective vaccines to prevent TB; using the Chemoprophylaxis for incubation patients to prevent their conversion into active TB; raising people’s awareness of the prevention and control of TB and treatment after illness; isolating the infected to reduce the spread of TB. According to the latest report in the announcement that came at the first WHO Global Ministerial Conference on Ending tuberculosis in the Sustainable Development Era, we predict that it is challenging to control TB by 2030.

The hydrodynamic prediction model improved design efficiency of the flow control application. A global sampling process with space filling method, adaptive sampling method, and neural network is proposed to generate a hydrodynamic prediction model for the flow control device. The optimized Latin hypercube sampling method is applied to create a uniform set of initial sample points for analysis. An automatic computational fluid dynamics analysis process is developed to provide data for the hydrodynamic objective of the multi-stage throttle orifice plate. By self-learning from previous sample points, the new sample points placed in the region of interest are collected by the adaptive sampling method in several iterations. Two numerical examples and the flow rate modeling for the throttle orifice plate are provided to demonstrate the approximation capability of the proposed process. Ten prediction model cases are established by a back-propagation feed-forward neural network, and the test data show that the model constructed by optimized Latin hypercube sampling and adaptive sampling has 1% mean relative error and 4.85% maximum relative error, which is nearly 25% more accurate than the optimized Latin hypercube sampling model. The proposed global sampling process is efficient to build the hydrodynamic prediction model and shows potential in the flow control design area.

In probabilistic risk assessment (PRA), two main methods exist for quantifying fault trees: theoretical and empirical (sampling). The efficiency of PRA quantification varies depending on the sampling method used. This study evaluated the feasibility of using quasi-Monte Carlo simulation (Quasi-MCS) and progressive Latin hypercube sampling (P-LHS) for PRA quantification. Eight risk outcomes were derived through PRA for internal and external events in four cases. The PRA convergence, variability, and error rates of each sampling method were compared and analyzed. The comparison analysis revealed that all sampling methods had an error rate of approximately 2% with 9,000 total samples. P-LHS exhibited the best convergence and variability among the methods, followed by Quasi-MCS and LHS. Although Quasi-MCS showed more significant variability than LHS as the number of events increased, its error rate remained within 2% with 9,000 samples. Therefore, both P-LHS and Quasi-MCS are feasible for PRA quantification.

As one of the critical technologies for performance evaluation of automated vehicles, the test scenario generation has been widespread concerned. In this paper, we propose a novel test scenario generation technology based on optimized Latin Hypercube Sampling (OLHS) and Test Matrix method (TM), named HIS-MPSO, which is efficient to generate the test scenario that consider the complexity, coverage, and potential relationships of factors. Based on naturalistic driving data, numerous car-following scenarios are generated by HIS-MPSO. Then, an adaptive cruise control system (ACC) are evaluated in terms of the tracking errors, comfort, and safety using the generated scenarios. Results show that compared with other existing OLHS algorithms, the HIS-MPSO can better restore the relationships among test factors existed in realistic traffic scenarios.

This study proposes a deep learning (DL)-based method for identifying the parameters of equivalent circuit models (ECMs) for lithium-ion batteries using time-series voltage response data from current pulse charge–discharge experiments. The application of DL techniques to this task is presented for the first time. The best-performing baseline model among the recurrent neural network, long short-term memory, and gated recurrent unit achieved a mean absolute percentage error (MAPE) of 0.52073 across the five parameters. Furthermore, more advanced models, including a one-dimensional convolutional neural network (1DCNN) and temporal convolutional networks, were developed using full factorial design (FFD), resulting in substantial MAPE improvements of 37.8% and 30.4%, respectively. The effectiveness of Latin hypercube sampling (LHS) for training data generation was also investigated, showing that it achieved comparable or better performance than FFD with only two-thirds of the training samples. Specifically, the 1DCNN model with LHS sampling achieved the best overall performance, with an average MAPE of 0.237409. These results highlight the potential of DL models combined with efficient sampling strategies.