Optimal Input Research Articles

Optimizing carbon source addition to control surplus sludge yield via machine learning-based interpretable ensemble model.

Appropriate carbon source addition can save operational costs and reduce surplus sludge yield in the wastewater treatment plant (WWTP). However, the link between carbon source and surplus sludge yield remains neglected although machine learning (ML) has become a powerful tool for WWTP, and is a challenge due to more complex multidimensional pattern recognition. Herein, weighted average ensemble strategy was conducted to assemble multiple diverse basic models to obtain better prediction capability to optimize carbon source addition (Model-1) and further control surplus sludge yield (Model-2). The ensemble models significantly outperformed all single models with MAE of 5.82g/m3, MSE of 60.59 and R2 value of 0.98 in Model-1 and MAE of 15.09g/m3, MSE of 449.01 and R2 value of 0.93 in Model-2. The optimal input feature subset was explored to reduce model complexity, indicating that the final ensemble models can predict with high precision using relatively few features with MAE of 6.41g/m3, MSE of 78.49 and R2 value of 0.97 in Model-1 and MAE of 12.82g/m3, MSE of 232.71 and R2 value of 0.95 in Model-2. Furthermore, the final models were deployed into an offline web application to facilitate their utility in real-world settings, demonstrating 47.25 % savings in carbon source addition and 15.89 % reductions in surplus sludge yield for an extra month of running. This work offers an efficient approach for the WWTP to optimize carbon source addition and provides new insights into controlling surplus sludge yield.

Optimizing laser cutting of stainless steel using latin hypercube sampling and neural networks

Optimizing cutting parameters in fiber laser cutting of austenitic stainless steel is challenging due to the complex interplay of multiple variables and quality metrics. To solve this problem, Latin hypercube sampling was used to ensure a comprehensive and efficient exploration of the parameter space with a smaller number of trials (185), coupled with feedforward neural networks for predictive modeling. The networks were trained with a leave-one-out cross-validation strategy to mitigate overfitting. Different configurations of hidden layers, neurons, and training functions were used. The approach was focused on minimizing dross and roughness on both the top and bottom areas of the cut surfaces. During the testing phase, an average MSE of 0.063 and an average MAPE of 4.68% were achieved by the models. Additionally, an experimental test was performed on the best parameter settings predicted by the models. Initial modelling was conducted for each quality metric individually, resulting in an average percentage difference of 1.37% between predicted and actual results. Grid search was also performed to determine an optimal input parameter set for all outputs, with predictions achieving an average accuracy of 98.34%. Experimental validation confirmed the accuracy and robustness of the model predictions, demonstrating the effectiveness of the methodology in optimizing multiple parameters of complex laser cutting processes.

Study of the Straw Compaction Degree as a Function of Moisture Content, Particle Size and Process Temperature.

The paper presents research on the process of densifying rye-wheat straw for its use in producing mouldable biofuel. The straw used in the research is a waste material from a farm located in Wielkopolska, resulting from the cultivation of triticale for the purpose of producing feed for pig farming. The aim of the study is to determine the utilisation of this material for the production of an agglomerate for energy purposes, such as heating the farm's infrastructure. The research was conducted for two moisture levels of straw: M = 10% and 30%. Before the experiment, the straw was cut into particles of the desired size: S = 10, 20, 30, 40, 50 and 60 mm. The densification process was carried out at temperatures T = 25, 50, 100, 150 and 200 °C, subjecting the straw to a compaction pressure of 15 MPa. Based on experimental studies, two values of the densification degree were determined: x1-the densification degree under load; and x2-the densification degree after unloading. The densification degree x2 is more relevant from the perspective of storage and transport. ANOVA analysis of the results showed that the most significant factors affecting x1 were particle size S and process temperature T, with higher x1 values obtained for straw moisture of 30%. The ANOVA analysis of the densification degree after unloading (x2) revealed that higher x2 values were achieved for straw with 10% moisture and the smallest particle size of 10 mm. The most significant factors affecting x2 were particle size and moisture content. Studies of the friction coefficient between the straw and the materials of the densification equipment components indicated that the optimal process temperature is 150 °C. The conducted research and the obtained results determined the optimal input parameters for the process and also provided a solid support for further studies, including investigation of the influence of other factors, such as binders.

Optimal Periodic Impulsive Control for Orbital Stabilization of Underactuated Systems

Abstract The orbital stabilization problem for underactuated systems with a single passive degree-of-freedom is revisited. The impulse controlled Poincaré map (ICPM) approach, in which stabilizing impulsive inputs are applied on a Poincaré section, has distinct advantages over existing methods but feedback compensation once every oscillation limits the rate of convergence to the desired orbit. To overcome these limitations, we propose stabilization through application of multiple impulsive inputs during each oscillation. An optimal control problem is formulated to minimize a quadratic cost functional and the optimal inputs are obtained by solving a discrete periodic Riccati equation. Simulation results for a Pendubot are presented, highlighting the advantages of the control design over the ICPM method in terms of convergence rate and robustness to parameter uncertainty.

Enhancing ride comfort and driving safety in urban autonomous buses: A data-driven hierarchical longitudinal motion planning using approximate and stochastic MPCs

This paper presents an advanced hierarchical longitudinal motion planning strategy for autonomous buses (ABs), specifically designed to enhance ride comfort and ensure driving safety – key challenges in contemporary autonomous vehicle technologies. ABs typically face limitations due to a narrower perception range and heterogeneous motion profiles compared to human drivers. To tackle these issues, our study employs a dual-phase motion planning framework utilizing both approximate and stochastic Model Predictive Controls (MPCs). In the long time-horizon planning phase, we aim to enhance ride comfort by optimizing a reference motion profile that covers the full perception range of the ABs. This optimization is facilitated by leveraging deep neural networks to approximate the MPC policy, ensuring efficient real-time computational performance. The short time-horizon planning phase focuses on driving safety by determining optimal control inputs, considering vehicle dynamics, and integrating safety-related chance constraints with stochastic MPC. Additionally, this phase fine-tunes the weights in the MPC’s objective function based on risk indices derived from human driving data, balancing the priorities of safety in hazardous situations and maintaining ride comfort. Our hierarchical motion planning strategy ensures both consistent ride comfort and safety in critical scenarios. Through comprehensive evaluations using computer simulations and vehicle tests in actual urban bus-only lanes, the proposed method has proven to significantly enhance ride comfort while maintaining driving safety. This approach contributes to the ongoing improvement of the operational quality of ABs, supporting public transportation goals and meeting passenger expectations effectively.

Machine learning for real-time reservoir operation simulation: comparing input variables and algorithms for the Sirikit Reservoir, Thailand

ABSTRACT Machine learning (ML) models offer advantages over process-based models for real-time reservoir operation modelling, yet the impact of input variable selection (IVS) and data pre-processing on model performance remains underexplored. This study investigates various input variables for simulating daily reservoir outflow, using the Sirikit reservoir in Thailand as a case study. The datasets include daily Sirikit storage and inflow, outflow of Bhumibol (neighbouring reservoir), downstream discharge, and temporal factors (month and day of the week). Time series decomposition and correlation analyses were used to assess data relationships. We tested seven ML models: multiple linear regression, support vector machine, K-nearest neighbour, classification and regression tree, random forest, multi-layer perceptron, and recurrent neural network (RNN). The optimal input set comprised the previous day’s storage, inflow from 2 days before to 2 days after, and month. With these inputs, all ML models simulated outflow adequately (KGEtraining = 0.42–1.0 and KGEtesting = 0.46–0.56), with RNN showing the most potential for improvement. Input scaling significantly enhanced model performance, reducing RMSEtraining by 44 m3 s-1 and RMSEtesting by 14 m3 s-1. This study’s novelty lies in its comprehensive insights of IVS and data scaling, highlighting their critical roles in enhancing ML model application for operational reservoir simulations.

Magnesium Alloy and Silicon Nitrides Composite Study: Synthesis and Abrasive Water Jet Machining Behavior and Optimization

<div>Related to traditional engineering materials, magnesium alloy-based composites have the potential for automobile applications and exhibit superior specific mechanical behavior. This study aims to synthesize the magnesium alloy (AZ61) composite configured with 0 wt%, 4 wt%, 8 wt%, and 12 wt% of silicon nitride micron particles, developed through a two-step stir-casting process under an argon environment. The synthesized cast AZ61 alloy matrix and its alloy embedded with 4 wt%, 8 wt%, and 12 wt% of Si<sub>3</sub>N<sub>4</sub> are subjected to an abrasive water jet drilling/machining (AJWM) process under varied input sources such as the diameter of the drill (D), transverse speed rate (v), and composition of AZ61 composite sample. Influences of AJWM input sources on metal removal rate (MRR) and surface roughness (Ra) are calculated for identifying the optimum input source factors to attain the best output responses like maximum MRR and minimum Ra via analysis of variant (ANOVA) Taguchi route with L16 design approach. The ANOVA analysis revealed that D, v and the composition of AZ61 alloy composite contribute 26.45%, 16.28%, and 20.84%, respectively, to the output response conditions for higher MRR. Additionally, design 7 exhibits a high MRR of 0.017 g/s and a surface roughness (Ra) of 0.84 μm. The optimum AWJM input source of design 7 is proposed for industries to mass production applications.</div>

Estimating production functions using costs when outputs are restricted

Conventional proxy-variable approaches to production-function identification assume that output quantities are flexibly adjustable, with firms endogenously choosing them in pursuit of profits. But in not so few industries outputs are exogenously restricted, constraining input allocations by producers. Model assumptions of most proxy estimators become untenable in such settings. We provide a new structural methodology based on a dual formulation of firm production, indirectly identifying technology and productivity from cost data while appropriately accommodating exogenous output restrictions. We showcase our methodology on a panel of U.S. natural gas-fired power-generation plants in 2005–2017 facing price-inelastic demand that restricts their outputs in the short run. We find that power plants were scale-efficient, with (short-run) marginal costs of net electricity generation declining over time. With little evidence of productivity growth in the industry during our sample period, we conclude that the reduction in marginal costs was fueled mainly by diminishing input prices and, potentially, convergence to the optimal long-run input allocation.

Data envelopment analysis and multi-objective genetic algorithm based optimization of energy consumption and greenhouse gas emissions in rice-wheat system

This study seeks to optimize energy use and reduce greenhouse gas emissions in rice-wheat cropping systems across Chhattisgarh, Bihar, and Punjab, India, using Data Envelopment Analysis and a multi-objective genetic algorithm. Energy inputs, including labour, machinery, diesel, fertilizers, herbicides, pesticides, fungicides, and irrigation, were evaluated across 65 farms. The average energy input was 39,706 ± 4877 MJ ha⁻1, while the average output was 140,961 MJ ha⁻1. The highest energy expenditures were attributed to nitrogen (15,995 ± 2973 MJ ha⁻1), diesel (5978 ± 358 MJ ha⁻1), and machinery (4438 ± 141 MJ ha⁻1). Data envelopment analysis results indicated that 26.15 % of farms operated at technical efficiency, with an average technical efficiency score of 0.833 and scale efficiency of 0.838. The Banker, Charnes, and Cooper model suggested an optimal energy input of 24,635 MJ ha⁻1. multi-objective genetic algorithm further optimized energy use, achieving a reduction of 13,440 MJ ha−1 compared to data envelopment analysis results alone. Conventional farming systems emitted 67,410 kg CO2-eq ha⁻1, while optimized farms achieved a reduction of 34 kg CO2-eq ha⁻1. These findings highlight the potential for substantial energy savings and greenhouse gas reductions through optimized input management, promoting more sustainable agricultural practices by minimizing reliance on chemical fertilizers, diesel, and machinery in rice-wheat systems.

Experimental study of data-driven model predictive control on transcritical CO2 thermal system in electric vehicles

The transcritical CO2 thermal system has been considered an effective and completable candidate for providing space/battery cooling and heating in electric vehicles. For a realistic system, the optimal performance relies on an effective control strategy. This paper presents a model predictive control approach to optimize the real-time operation of the transcritical CO2 thermal system and conduct a complete experimental investigation. A data-driven control-oriented model is first developed to predict the next steps in system behaviours in a finite time domain. The model predictive controller is designed to provide the optimal inputs based on the control-oriented model and the designed objective function, in which optimal system COP can be achieved provided that the cooling/heating capacity is maintained. Then, a complete test rig is built in a psychrometric test room to experimentally investigate the operating performance using the proposed model predictive control strategy. The experiments are conducted under fixed and variable ambient temperatures for both cooling and heating conditions. The experimental results indicate that the model predictive control strategy can accurately forecast system states and determine optimal control inputs for the transcritical CO2 thermal system to achieve the highest operating COP with the required cooling/heating capacity in electric vehicles.

NN-Based Reinforcement Learning Optimal Control for Inequality-Constrained Nonlinear Discrete-Time Systems With Disturbances.

Based on actor-critic neural networks (NNs), an optimal controller is proposed for solving the constrained control problem of an affine nonlinear discrete-time system with disturbances. The actor NNs provide the control signals and the critic NNs work as the performance indicators of the controller. By converting the original state constraints into new input constraints and state constraints, the penalty functions are introduced into the cost function, and then the constrained optimal control problem is transformed into an unconstrained one. Further, the relationship between the optimal control input and worst-case disturbance is obtained using the Game theory. With Lyapunov stability theory, the control signals are ensured to be uniformly ultimately bounded (UUB). Finally, the effectiveness of the control algorithms is tested through a numeral simulation using a third-order dynamic system.

Fourier Hilbert: The Input Transformation to Enhance CNN Models for Speech Emotion Recognition

Signal processing in general, and speech emotion recognition in particular, have long been familiar Artificial Intelligence (AI) tasks. With the explosion of deep learning, CNN models are used more frequently, accompanied by the emergence of many signal transformations. However, these methods often require significant hardware and runtime. In an effort to address these issues, we analyze and learn from existing transformations, leading us to propose a new method: Fourier Hilbert Transformation (FHT). In general, this method applies the Hilbert curve to Fourier images. The resulting images are small and dense, which is a shape well-suited to the CNN architecture. Additionally, the better distribution of information on the image allows the filters to fully utilize their power. These points support the argument that FHT provides an optimal input for CNN. Experiments conducted on popular datasets yielded promising results. FHT saves a large amount of hardware usage and runtime while maintaining high performance, even offers greater stability compared to existing methods. This opens up opportunities for deploying signal processing tasks on real-time systems with limited hardware.

Fizzy extraction secondary electrospray ionization mass spectrometry for analysis of volatile solutes in high-matrix samples

Fizzy extraction (FE) is a rapid way to prepare liquid samples containing volatile organic compounds (VOCs) for analysis. The sample is first infused with a carrier gas, and subsequent rapid decompression of the sample chamber causes it to effervesce spontaneously. Secondary electrospray ionization (SESI) enables analysis of VOCs using mass spectrometry (MS). Here, we report direct hyphenation of FE with SESI-MS. In order to adjust the flow rate of the gaseous extract to match the optimum input flow rate of the ion source, flow was split with a tee junction. The following parameters have been optimized: pressure inside the sample chamber (during saturation), high voltage applied to the nano-ESI solution, temperature of desolvation line, temperature of heated block, temperature of the SESI chamber, gas pressure applied to the headspace of nano-ESI solution, and length of tubing connected to the tee junction. The optimum parameters were 150 kPa, +4.0 kV, 225 °C, 200 °C, 90 °C, 18 kPa, and 6 cm, respectively. The developed method was characterized using MS and high-speed imaging. The limits of detection for ethyl butyrate, ethyl pentanoate, ethyl hexanoate, ethyl octanoate, and ethyl decanoate are 1.11 × 10−7 M, 4.33 × 10−7 M, 8.13 × 10−8 M, 1.09 × 10−7 M, and 2.19 × 10−7 M, respectively. The analysis time was only 2.5 min per sample (not counting the cleaning step). The method was further applied for profiling VOCs in selected dairy products. Interestingly, the FE-SESI-MS method is much more tolerant to the sample matrix (milk) than headspace vapor analysis using the same ion source.

Determination of creep function coefficients of viscoelastic pipes using a transient-guided machine learning model

ABSTRACT This study introduces a novel framework for the estimation of creep function coefficients in viscoelastic pipelines, employing a combination of machine learning (ML) and transient hydraulics. Specifically, the efficacy of eXtreme Gradient Boosting (XGBoost) as an advanced regression method is evaluated within this framework. A transient simulation model, utilizing the method of characteristics, is formulated to create the reservoir-pipe-valve (RPV) scenarios under diverse boundary conditions. After conducting model calibration, the model is used to generate datasets with the transient hydraulic responses at the measurement points. The fast Fourier transform (FFT) is then applied to transform the generated samples into the frequency domain. Feature selection is accomplished through principal component analysis (PCA) to identify optimal input variables for XGBoost. In estimating the creep function, six coefficients are employed, with the feature selection analysis indicating that each coefficient is associated with a specific signal. Importantly, it is demonstrated that attempting to estimate all six coefficients using a single set of signals is unfeasible. The results affirm the accuracy of the proposed ML-based framework in determining creep function coefficients.

The Role of Agricultural Inputs on Crop Productivity: The Case of Zigam Woreda

In spite of the importance of agriculture in Ethiopia, it is characterized by low productivity and has been unable to produce sufficient quantities of output to feed the country’s population. In light of this various development strategies has been undertaken to improve the performance of agriculture. Intensification of agriculture through the use of new agricultural technologies has been emphasized over the last three decades. This study attempted to examine the contribution of agricultural input for crop productivity. The data for the study was collected from 91 sample farmers. This study was the study used both primary and secondary data. In this study researcher was used simple random sampling techniques. This research was use cross sectional approach and econometric method of data analysis to investigate the role of age, sex, land size, labor force, fertilizer, improved seed, extension service, and access to credit, education level and pesticides for crop production by collecting data from the household. In econometric method of data analysis researcher was used ordinary least square (OLS) Model. The econometric result show that land size, labor force, improved seed, fertilizer, credit service, extension service and education level have positive and significant effect on crop production. However, pesticide has a negative and significant impact on crop production. From the explanatory variables, education level has a higher coefficient. This indicates education level is more significant for crop production. According regression result <i>R</i><sup><i>2</i></sup> is 0.97, which implies 97% of output function is explained by the selected ten (10) explanatory variables. The policy implication is that to reduce farmers resistant to use farm inputs and to create knowledge about the optimal input use educate and training of farmers is necessary.

Multi‐Response Optimization of an Orthotropic Steel Deck Section with Thermo‐mechanical Tensioning

In this investigation, Taguchi's grey relational analysis (GRA) is utilized to optimize parameters in an orthotropic steel deck section subjected to thermo‐mechanical tensioning. In this study, it is aimed to determine optimal values for deck thickness, welding sequence, and welding speed, which are evaluated through finite‐element analysis based on outputs such as longitudinal stress, transverse stress, and deflection. In this study, a previously validated numerical simulation model is extended by conducting a parametric analysis to investigate the influence of varying parameters on the model's behavior. An analysis of variance is conducted to identify the most significant variable affecting the input parameters. Additionally, a confirmatory test is also performed to validate the characteristics of the Grey Relational Grade. In the results, it is indicated that faster welding speeds result in increased deflection of the deck plate. Furthermore, the transition from compressive to tensile longitudinal stress is more abrupt in thinner decks (9–10 mm) compared to thicker decks (11–12 mm). This redistribution of residual stresses is attributed to the increasing thickness of the deck plate. Optimal input levels (A2–B1–C1: 10 mm deck thickness, welding sequence case 1, and 0.005 m s−1 welding speed) are determined using GRA. Welding speed contributes the maximum to the optimization results among the evaluated variables.

Learning‐based tracking control of AUV: Mixed policy improvement and game‐based disturbance rejection

AbstractA mixed adaptive dynamic programming (ADP) scheme based on zero‐sum game theory is developed to address optimal control problems of autonomous underwater vehicle (AUV) systems subject to disturbances and safe constraints. By combining prior dynamic knowledge and actual sampled data, the proposed approach effectively mitigates the defect caused by the inaccurate dynamic model and significantly improves the training speed of the ADP algorithm. Initially, the dataset is enriched with sufficient reference data collected based on a nominal model without considering modelling bias. Also, the control object interacts with the real environment and continuously gathers adequate sampled data in the dataset. To comprehensively leverage the advantages of model‐based and model‐free methods during training, an adaptive tuning factor is introduced based on the dataset that possesses model‐referenced information and conforms to the distribution of the real‐world environment, which balances the influence of model‐based control law and data‐driven policy gradient on the direction of policy improvement. As a result, the proposed approach accelerates the learning speed compared to data‐driven methods, concurrently also enhancing the tracking performance in comparison to model‐based control methods. Moreover, the optimal control problem under disturbances is formulated as a zero‐sum game, and the actor‐critic‐disturbance framework is introduced to approximate the optimal control input, cost function, and disturbance policy, respectively. Furthermore, the convergence property of the proposed algorithm based on the value iteration method is analysed. Finally, an example of AUV path following based on the improved line‐of‐sight guidance is presented to demonstrate the effectiveness of the proposed method.

Prediction of feed force with machine learning algorithms in boring of AISI P20 plastic mold steel

Feed force plays a critical role in machining performance, including efficiency, hole quality, energy consumption, and tool life. While feed force is typically measured using dynamometers, these devices are often impractical in real-world experiments due to their complexity and cost. In particular, boring operations, which are essential for enlarging holes within tight tolerances, depend on precise feed force control to maintain high-quality hole outcomes. Achieving this precision requires reliable pre-process feed force estimation. However, traditional analytical methods for calculating feed force in boring operations often fall short, largely due to factors such as in-hole dynamics, the inclination of the rake surface, and the slenderness of the boring bar. This study aims to apply single and hybrid machine learning methods to accurately model and predict feed force in the boring process. Unlike previous research, which focused on feature selection, we propose the use of all possible combinations of input feature subsets to determine the optimal input features. The results indicate that feed force can be predicted effectively using these optimized feature combinations. Such accurate feed force prediction is crucial for effective process optimization, material selection, and process planning.

Requirements for Standard and I/O Cell Library Domestic Characterization System Development

The article is devoted to define requirements and discover specific featuresfor domestic standard cell and I/O libraries characterization EDA tool. A general structure of the library characterization system is presented. The basic requirements indicted for the system are considered and analyzed in the article. A comparative analysis of correspondence to these requirements for domestic and foreign characterization tools is done. The specific features the domestic library characterization system should satisfy are formulated. Among the main features the following are outlined in the article: compatibility and support of domestic OS, support of well-known foreign and domestic Spice-simulators, Russian language support for documentation, scalability, ensuring confidence of the output data. A brief review of typical tasks, intrinsic for characterization tool development and corresponding algorithms, required for its resolving is given. The following issues along with appropriate proposals for its solution are mentioned: automated definition of optimal input waveforms for sequential cells initialization, searching for maximum capacitance load for the cell, setup and hold parameters calculation for sequential cells, method of optimal output current PWL approximation for CCS parameters calculation. As a conclusion of the investigation the current status of domestic characterization systems development is analyzed. In the outcome a problem of domestic library characterization tool that aligned with all the requirements and support the specific features defined above has been raised.

Prediction of formation fracture pressure based on reinforcement learning and XGBoost

Abstract Clearly determining the magnitude of fracture pressure is a crucial indicator for fracturing design. Traditional methods for predicting fracture pressure suffer from challenges such as difficulties in obtaining required data, low prediction accuracy, and local limitations in application. In light of these issues, the article proposes a fracture pressure prediction model based on reinforcement learning and XGBoost utilizing geophysical well logging data. Based on the relevance analysis, optimal input parameters, including DEPTH, DEN, AC, GR, CRL, and RT, are selected from geophysical well logging data. We have developed a framework for a fracture pressure prediction model based on XGBoost, wherein hyperparameters are fine-tuned using an improved Q-learning algorithm. The optimized XGBoost model for fracture pressure prediction attains outstanding performance metrics, including an R 2 value of 0.992, a root mean square error of 0.006%, and a mean absolute error of 0.539%. In direct comparison with grid search, Bayesian optimization, and ant colony optimization, the improved Q-learning algorithm emerges as the most effective optimization approach. The predictions generated by the proposed method exhibit remarkable consistency with fracture pressure data measured on-site. This approach successfully addresses the shortcomings encountered with traditional fracture pressure prediction methods, such as inadequate accuracy, demanding data prerequisites, and constrained applicability.