Integrated Optimization Algorithm Research Articles

Harnessing Educational Big Data Analytics for Decision-Making in Enhancing School Teaching Quality

The application of educational big data analytics holds significant importance in enhancing decision-making processes for school teaching quality. This study explores the effective utilization of educational big data analytics technologies to support the improvement of teaching quality in schools. Initially, the challenges and needs faced by current school teaching quality decision-making were analyzed, highlighting the critical role of educational big data analytics in this context. Subsequently, the limitations and gaps in existing study were identified through a review of related studies, underscoring the study value of this study. Based on this foundation, this study progresses through an examination of the decision-making factors that influence school teaching quality, problem description and model assumptions, construction of decision models, and model solutions using genetic algorithms. By analyzing key factors and constraints in the decision-making process for school teaching quality and integrating optimization algorithms, a viable decision support model was proposed and empirically analyzed. This study aims to provide a scientific basis for school administrators and decision-makers, thereby promoting continuous improvement in school teaching quality.

Scientific Integration of Operations Research and Machine Learning for Data Centre Optimization

In this study, we explore how to optimize data center operations by combining Operations Research (OR) and Machine Learning (ML) methodologies with Python-based categorization algorithms. Using Scikit-learn and TensorFlow, two Python libraries, we investigate how ML algorithms might be integrated with OR techniques like queuing theory and linear programming to forecast workloads and allocate resources more effectively. Problems with scheduling workloads, allocating resources, and managing energy consumption are at the heart of our research into data center optimization. The goal of this comprehensive framework is to create more effective and environmentally friendly data centre operations by systematically evaluating Python-based categorization models in response to changing workload demands and environmental circumstances. Introduction: The backbone of our digital infrastructure, data centers stand tall in the ever-changing world of contemporary technology. A vast variety of online services, including social media platforms, e-commerce websites, cloud computing, and big data analytics, rely on the servers, storage devices, networking gear, and other essential components housed in these expansive facilities. Meeting the ever-increasing demands for computational resources while simultaneously enhancing performance, efficiency, and cost-effectiveness is a daunting task for data centers, which are already struggling to keep up with the exponential growth in both the amount and complexity of digital data. Objectives: Our goal in writing this article is to delve into the ways in which data center optimization intersects with Operations Research and Machine Learning. Data center optimization presents a wide range of problems, and this course will explore the theory, methods, and best practices for using OR and ML to solve these problems. To develop an integrated framework that combines operations research (OR) and machine learning (ML) techniques to optimize the performance, energy efficiency, and reliability of data centers. Methods: Optimization strategies that improve data center operations in terms of performance, efficiency, and sustainability. These proposed strategies make use of both OL and ML techniques. Data center operators can optimize resource allocation, workload management, temperature control, energy consumption, and anomaly detection in real-time by formally stating the optimization problem in a mathematical framework. This allows for informed decision-making, systematic analysis of trade-offs, and the implementation of adaptive control strategies. Results: The visualization depicts the projected energy usage in terms of bandwidth for both approaches, compared to the actual values. In general, although both methods demonstrate potential, additional refinement and optimization may be necessary to attain superior outcomes in real-life situations. This discussion presents an analysis of the performance of both procedures and offers insights into their respective strengths and shortcomings, which can serve as a foundation for further investigation or improvement of the approaches. Conclusions: By using a comprehensive and multidisciplinary approach, we can optimize data centres in a way that boosts efficiency and performance while simultaneously encouraging innovation, resilience, and sustainability in data centre operations. Also, Data center gains in resource operators can achieve considerable utilization, energy efficiency, and overall system performance by integrating optimization algorithms, predictive analytics, and adaptive control strategies

Fine-tuning inflow prediction models: integrating optimization algorithms and TRMM data for enhanced accuracy.

This research explores machine learning algorithms for reservoir inflow prediction, including long short-term memory (LSTM), random forest (RF), and metaheuristic-optimized models. The impact of feature engineering techniques such as discrete wavelet transform (DWT) and XGBoost feature selection is investigated. LSTM shows promise, with LSTM-XGBoost exhibiting strong generalization from 179.81 m3/s RMSE (root mean square error) in training to 49.42 m3/s in testing. The RF-XGBoost and models incorporating DWT, like LSTM-DWT and RF-DWT, also perform well, underscoring the significance of feature engineering. Comparisons illustrate enhancements with DWT: LSTM and RF reduce training and testing RMSE substantially when using DWT. Metaheuristic models like MLP-ABC and LSSVR-PSO benefit from DWT as well, with the LSSVR-PSO-DWT model demonstrating excellent predictive accuracy, showing 133.97 m3/s RMSE in training and 47.08 m3/s RMSE in testing. This model synergistically combines LSSVR, PSO, and DWT, emerging as the top performers by effectively capturing intricate reservoir inflow patterns.

A novel machine learning-based artificial intelligence method for predicting the air pollution index PM2.5

Accurate prediction of the Particulate Matter 2.5 (PM2.5) plays a crucial role in the accurate management of air pollution and prevention of respiratory diseases. However, PM2.5 as a time series is extremely difficult to accurately predict. In this paper, a Hybrid Integration (HIG) algorithm that combines data pre-processing, time-series decomposition, signal decomposition, a prediction module, a matching strategy, and a hybrid integrated optimization algorithm is proposed. First, the optimal parameters for the four individual models were selected by integrating multiple evaluation perspectives. Additivity was then determined by Seasonal and Trend decomposition using LOWESS (STL), followed by refinement decomposition using signal decomposition. The four new sequences were reconstructed using Range Entropy (RangeEn) and mapped to the models. Additionally, Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) methods were optimized using the HIG algorithm. The results show that the HIG-RNN and HIG-LSTM are more advantageous than the ordinary method in terms of reasonable weight assignment. Finally, an innovative confusion test method was developed to test the stability of the prediction direction. To ensure generalizability, validation was performed using PM2.5 data from two regions of China. The results show that the method significantly improves the prediction performance and provides a powerful tool for policy formulation and management.

Speaker Recognition Using MFCC-BPNN-HHO

Speaker recognition plays a pivotal role in speech processing. This paper proposes an enhancement to the Backpropagation Neural Network (BPNN) by incorporating Harris Hawks Optimization (HHO) for weight optimization, and evaluates its performance compared to the standalone BPNN. Both methods employ Mel Frequency Cepstral Coefficients (MFCC) for feature extraction from input data. The study assesses the proposed system on a dataset comprising 10 speakers, with each providing 10 utterances. Results demonstrate that the integrated MFCC-BPNN-HHO approach outperforms the standalone BPNN, achieving enhanced accuracy in speaker recognition tasks. Specifically, the accuracy of the BPNN-HHO was found to be significantly higher than that of the BPNN alone, indicating the effectiveness of the HHO optimization technique in improving speaker recognition accuracy. This study underscores the potential of integrating optimization algorithms like HHO with BPNN to further refine speaker recognition systems and contribute to advancements in speech processing technology. This approach has promising applications in access control, identity verification and other security-related domains where biometric authentication is essential. Keywords—MFCC, BPNN, HHO

Joint optimal beam forming and resource allocation in intelligent reflecting surface aided wireless power transfer rate splitting multiple access system

SummaryThe intelligent reflecting surface (IRS) has recently become the most promising technology to achieve maximum beam‐forming gain with a simultaneous wireless information and power transfer (SWIPT) system. Many existing studies perform a single IRS deployment or utilize space division multiple access (SDMA) and non‐orthogonal multiple access (NOMA) schemes. However, the existing schemes face high time complexity and block signal transmission for larger indoor communications. Hence, this article introduces a dual IRS‐aided multiuser SWIPT system by implementing rate‐splitting multiple access (RSMA) systems. To enhance the minimum achievable rate and ensure the minimum harvesting energy (HE), this study proposes a joint successive convex approximation with an integrated optimization algorithm (SCA‐IOA). The proposed algorithm jointly optimizes the reflecting beam forming and power splitting (PS) coefficient of all the users effectively. The simulation results demonstrate the effectiveness of the proposed techniques and produce an outstanding performance for deploying dual IRS and implementing RSMA compared to traditional SDMA and NOMA systems.

Medical Image-Based Diagnosis Using a Hybrid Adaptive Neuro-Fuzzy Inferences System (ANFIS) Optimized by GA with a Deep Network Model for Features Extraction

Predicting diseases in the early stages is extremely important. By taking advantage of advances in deep learning and fuzzy logic techniques, a new model is proposed in this paper for disease evaluation depending on the adaptive neuro-fuzzy inference system (ANFIS) with a genetic algorithm (GA) for classification, and the pre-trained DenseNet-201 model for feature extraction, in addition to the whale optimization algorithm (WOA) for feature selection. Two medical databases (chest X-ray and MRI brain tumor) for the diagnosis of two disease types were used as input in the suggested model. The optimization of ANFIS parameters was performed by GA to achieve the optimum prediction capability. DenseNet-201 for feature extraction was employed to obtain better classification accuracy. Having more features sometimes leads to lower accuracy, and this issue can be rectified using a feature selection strategy WOA which gave good results. The proposed model was evaluated utilizing statistical metrics root mean square error (RMSE), mean square error (MSE), standard deviation (STD), and coefficient of determination (R2), and it was compared with the conventional ANFIS model, with the proposed model (ANFIS-GA) showing a superior prediction capability over the ANFIS model. As a result, it can be concluded that the proposed ANFIS-GA model is efficient and has the potential for a robust diseases evaluation with good accuracy. Also, we conclude from this work that integrating optimization algorithms with ANFIS boosts its performance, resulting in a more accurate and reliable model.

Optimization of State of the Art Fuzzy-Based Machine Learning Techniques for Total Dissolved Solids Prediction

Total dissolved solid prediction is an important factor which can support the early warning of water pollution, especially in the areas exposed to a mixture of pollutants. In this study, a new fuzzy-based intelligent system was developed, due to the uncertainty of the TDS time series data, by integrating optimization algorithms. Monthly-timescale water quality parameters data from nearly four decades (1974–2016), recorded over two gaging stations in coastal Iran, were used for the analysis. For model implementation, the current research aims to model the TDS parameter in a river system by using relevant biochemical parameters such as Ca, Mg, Na, and HCO3. To produce more compact networks along with the model’s generalization, a hybrid model which integrates a fuzzy-based intelligent system with the grasshopper optimization algorithm, NF-GMDH-GOA, is proposed for the prediction of the monthly TDS, and the prediction results are compared with five standalone and hybrid machine learning techniques. Results show that the proposed integrated NF-GMDH-GOA was able to provide an algorithmically informed simulation (NSE = 0.970 for Rig-Cheshmeh and NSE = 0.94 Soleyman Tangeh) of the dynamics of TDS records comparable to the artificial neural network, extreme learning machine, adaptive neuro fuzzy inference system, GMDH, and NF-GMDH-PSO models. According to the results of sensitivity analysis, Sodium in natural bodies of water with maximum value of error (RMSE = 56.4) had the highest influence on the TDS prediction for both stations, and Mg with RMSE = 43.251 stood second. The results of the Wilcoxon signed rank tests also indicated that the model’s prediction means were different, as the p value calculated for the models was less than the standard significance level (α=0.05).

Multi-parameter optimization for Vis–NIR spectroscopic analysis of multiple indicators of soil heavy metal in the tideland reclamation area of the Pearl River Delta

ABSTRACT Using visible and near-infrared (Vis–NIR) spectroscopy combined with novel chemometric methods, the rapid reagent-free simultaneous analysis model for Cu, Zn, Ni, and Cr contents in tideland reclamation soil in the Pearl River Delta was established. Based on Savitzky–Golay (SG) smoothing and partial least squares (PLS) regression, a multi-parameter optimization platform (SG-PLS) covering 264 modes was constructed to select appropriate spectral preprocessing mode for each indicator. The equidistant combination PLS (EC-PLS) method with three cycle parameters was adopted for the first large-scale screening of wavelength models. In addition, wavelength phase-out PLS (WSP-PLS) and repetition rate priority combination methods were used as the secondary optimization method. The well-executed competitive adaptive reweighted sampling PLS was also used for comparison. The validation samples that were not involved in modeling were used to validate the selected three group models. For the four indicators, the method of EC-PLS combined with WSP-PLS achieved the best validation effect. In validation, the root mean square error (SEPV), relative root mean square error (RSEPV), and correlation coefficients (RP,V) of prediction were 2.46 mg kg−1, 4.5%, and 0.959 for Cu; 51.37 mg kg−1, 24.6%, and 0.900 for Zn; 3.59 mg kg−1, 9.1%, and 0.739 for Ni; and 9.15 mg kg−1, 8.4%, and 0.843 for Cr, respectively. Results indicated that low relative error and high prediction correlation confirmed the feasibility of using Vis–NIR spectroscopy to analyze soil heavy metal contents. The proposed multi-parameter, multi-stage integrated optimization algorithm could be applied to a wider field of spectral analysis.

Weighted mean of vectors optimization algorithm and its application in designing the power system stabilizer

Accurate design of the power system stabilizer (PSS) models is a crucial issue due to their significant impact on the stability of power system operation. However, identifying the parameters of a PSS model is a challenging task owing to its nonlinearity and multi-modality characteristics. Due to such characteristics, handling algorithms may be prone to stagnation in local optima. Therefore, this paper proposes a potent integrated optimization algorithm by comprising the weIghted meaN oF vectOrs (INFO) optimizer with chaotic-orthogonal based learning (COBL) and Gaussian bare-bones (GBB) strategies, named INFO-GBB, for achieving the optimal parameters of a PSS model used in a single-machine infinite-bus (SMIB) system. In the INFO-GBB, the COBL aims to enhance the searching capability to explore new regions using the orthogonal design aspect and thus improving the diversity of solutions. Also, the GBB is adopted to assist the algorithm to perform an immediate vicinity of the best solution and thus enhances the exploitation capabilities. The effectiveness and efficacy of the INFO-GBB algorithm is validated on CEC 2020 benchmark suits and the designing task of the PSS model. The achieved results by the INFO-GBB are compared with eighteen well-known algorithms. The statistical verifications along with the Friedman test have ascertained that the INFO-GBB is capable of achieving promising performances compared to the other counterparts. The results obtained based on the Friedman test illustrate that the INFO-GBB offers superior performance over the state-of-the-art algorithms as it outperforms fifteen out of eighteen algorithms by an average rank greater than 61% for benchmark problems while outperforming O-LSHADE, LSHADE, and TSA algorithms by 25%,33%, and 58%, respectively. Furthermore, the applicability of the INFO-GBB is realized through designing the PSS model used in a SMIB system. The obtained results indicate that the INFO-GBB algorithm exhibits accurate and superior performance compared to other peers as it provides the lowest value for the integral of time multiplied absolute error (ITAE) performance index which is used as an objective function. For example, the achieved results of the mean ITAE found by INFO-GBB is 1.36E−03 with improvement percentages of 24.93%, 19.78%, 13.04%, 26.64%, and 24.86%, over the LSHADE, GWO, EO, RSA, and original INFO algorithms, respectively. Therefore, the INFO-GBB can efficiently affirm its superiority and stability to deal with the function optimization task and parameters’ estimation of the PSS model.

An adapter-based architecture for evaluating candidate solutions in energy system scheduling

Increasing shares of volatile generation and non-steerable demand raise the need for automated control of the Energy Systems (ESs). Various solutions for management and schedule-based control of energy facilities exist today. However, the amount and diversity of applications lead to a multitude of different automated energy management solutions. Different optimization algorithms have proven more or less effective for energy management. The multitude of optimization algorithms and energy management solutions require flexible, modular, and scalable integrations. We present a novel Optimization Service (OS) for easily integrating optimization algorithms while evaluating candidate solutions in the context of ESs applications. We propose an adapter-based architecture using metadata and domain knowledge to bridge between clients, e.g. smart grid applications and optimization algorithms. The architecture interfaces different clients with optimizers in a flexible and modular way. The clients provide metadata-based descriptions of optimization jobs translated by OS. OS then interacts with optimizers and evaluates candidate solutions. A consistent definition of interfaces for clients and optimization algorithms facilitates the modular evaluation of candidate solutions. OS’s separation of client and optimization algorithms increases scalability by managing computational resources independently. We evaluate the presented architecture for scheduling a so-called Energy Hub (EH) as a test case describing a simulation scenario of a renewable EH embedded in grid scenarios from an industrial area in Karlsruhe, Germany. OS utilizes an Evolutionary Algorithm (EA) to optimize schedules for cost and strain on the electrical grid. The use case exemplifies OS’s advantages in a proof-of-concept evaluation.

Integrated Optimization of Hybrid Steam-Solvent Processes in a Post-CHOPS Reservoir with Consideration of Wormhole Networks

Abstract In this paper, an integrated technique has been developed to evaluate and optimize performance of hybrid steam-solvent processes in a post-cold heavy oil production with sand (CHOPS) reservoir with consideration of wormhole networks. A reservoir geological model is developed and calibrated by history matching reservoir pressure with oil, gas, and water production rates as the input constraints, while its wormhole network is characterized with a newly developed pressure-gradient-based (PGB) sand failure criterion conditioned to sand production. Once calibrated, the reservoir geological model incorporated with the wormhole network is then employed to evaluate and optimize performance of hybrid steam-solvent processes under various conditions, during which the net present value (NPV) is maximized with an integrated optimization algorithm by taking injection time, soaking time, production time, and injected fluid composition as controlling variables. It is found that a huff-n-puff process imposes a positive impact on enhancing oil recovery when wormhole network is fully generated and propagated. Addition of alkane solvents into CO2 stream leads to a higher oil recovery compared with that of the CO2 only method, while all hybrid steam-solvent injection achieve high oil recovery by taking advantage of both thermal energy and solvent dissolution. It is found that the NPV reaches its maximum value when the steam temperature is 200 °C for the optimized hybrid steam-solvent scenario.

An effective integration optimization algorithm for regional fracturing design and drilling placement

Hydraulic fracturing horizontal well technology has a pivotal role in exploiting unconventional resources, but it often requires considerable investment. Integrated optimization of the fracturing design, horizontal well, and drilling platform placement in the whole region is an efficient way to reduce cost and increase economic benefits. However, it is difficult to quickly complete these optimizations because traditional reservoir numerical simulation requires extensive resources in terms of labor, time, and computation. Therefore, we proposed a new algorithm, named Gaussian Process Regression-Bilevel Programming Genetic-Critical Circle Intersection Momentum Shift (GPR-BPG-CCIMS) algorithm to effectively co-optimize the fracturing design, horizontal well and drilling platform placement in the whole region. BPG algorithm shares the merits of Bilevel Programming (BP) and genetic algorithm (GA), which is used for optimizing the horizontal well and fracturing parameters. CCIMS algorithm is applied to optimize the drilling platform placement by merging and shifting drilling platform positions. GPR model is a production proxy model based on machine learning to improve optimization efficiency. The proposed algorithm was applied to a real shale gas reservoir to verify the effectiveness of the algorithm. The results showed that the GPR model exhibited great prediction accuracy by providing the high coefficient of determination (R2 = 0.999), the high percentage of accuracy-precision (PAP = 99.017) and the low root mean square error (RMSE = 0.022). The optimal fracture half length, cluster spacing, fracture fluid per meter, horizontal section length, well location, and drilling platform location were obtained by GPR-BPG-CCIMS algorithm. Compared with the actual net present value (NPV), the optimized NPV was increased by 181 million dollars. The proposed algorithm successfully co-optimizes the engineering parameters of the whole region in a short consumption time, which effectively supports the development scheme design and decision-making. The optimization framework can be extended to other shale gas basins.

A Building Energy-Saving Method of Small High-Rise Office Building Based on BIM Model

Building green energy-saving is an important area of current architectural application research. Artificial intelligence algorithm can effectively improve the design of building energy-saving digital, intelligent level. Therefore, this paper proposes the use of BIM model based on genetic algorithms to optimize the application of building energy conservation research. In this paper, the basic structure of genetic algorithm and advantages and disadvantages of the algorithm are described. According to the need of optimal design of building energy-saving integration, multiobjective optimization function is introduced and a multiobjective optimization genetic algorithm is established to improve the prediction effect of building energy efficiency model. Finally, results of analysis of some of the influencing factors and the simulation test of building energy-saving integrated optimization algorithm in small high-rise office building show that the improved genetic algorithm can effectively improve the effect of energy-saving integrated optimization and has good application prospects.

Integrated optimization of fracture parameters for subdivision cutting fractured horizontal wells in shale oil reservoirs

Horizontal well with subdivision cutting fracturing technology is a new technical way to improve the efficient development of shale reservoirs. Subdivision cutting fracturing has the characteristics of stronger inter-fracture interference, more variable parameters and constraints. This fracturing technology has no efficient method to optimize the discrete and continuous fracture parameters integrally. A new automatic integrated optimization algorithm is proposed. The algorithm adopts the SPSA nested binary search algorithm to optimize the discrete and continuous fracture parameters. Considering the flow mechanism of multi-scale media in shale reservoirs, a numerical flow simulation model of subdivision cutting fractured horizontal wells is established. Then, taking the NPV as the objective function, a multi-parameter integrated optimization model is established. Make the method apply to a typical test shale reservoir. The optimization results show that the NPV of the optimum fracture parameter combination is 37.69%–57.36% higher than that of the uniform fracture distribution combination. And the optimum fracture combination presents a spindle-like distribution. This method is applied to an actual shale reservoir in Xinjiang to verify the broad application. The fracture parameter combination of two parallel horizontal wells in the reservoir is optimized. The optimization results show that the NPV has been dramatically improved, with an increasing range of 37.94%–96.84%. And the fractures of two adjacent horizontal wells present zipper staggered distribution. The integrated optimization method of fracture parameters can provide theoretical guidance for hydraulic fracturing design.

Identification of Hydrodynamic Dispersion Tensor by Optimization Algorithm Based LBM/CMA-ES Combination

The hydrodynamic dispersion tensor (HDT) of a porous medium is a key parameter in engineering and environmental sciences. Its knowledge allows for example, to accurately predict the propagation of a pollution front induced by a surface (or subsurface) flow. This paper proposes a new mathematical model based on inverse problem-solving techniques to identify the HDT (noted D=) of the studied porous medium. We then showed that in practice, this new model can be written in the form of an integrated optimization algorithm (IOA). The IOA is based on the numerical solution of the direct problem (which solves the convection–diffusion type transport equation) and the optimization of the error function between the simulated concentration field and that observed at the application site. The partial differential equations of the direct model were solved by high resolution of (Δx=Δy=1 m) Lattice Boltzmann Method (LBM) whose computational code is named HYDRODISP-LBM (HYDRO-DISpersion by LBM). As for the optimization step, we opted for the CMA-ES (Covariance Matrix Adaptation-Evolution Strategy) algorithm. Our choice for these two methods was motivated by their excellent performance proven in the abundant literature. The paper describes in detail the operation of the coupling of the two computer codes forming the IOA that we have named HYDRODISP-LBM/CMA-ES. Finally, the IOA was applied at the Beauvais experimental site to identify the HDT D=. The geological analyzes of this site showed that the tensor identified by the IOA is in perfect agreement with the characteristics of the geological formation of the site which are connected with the mixing processes of the latter.

Integrated optimization algorithm: A metaheuristic approach for complicated optimization

This paper proposes an integrated optimization algorithm (IOA) designed for solving complicated optimization problems that are non-convex, non-differentiable, non-continuous, or computationally intensive. IOA is synthesized from 5 sub-algorithms: follower search, leader search, wanderer search, crossover search, and role learning. The follower search finds better solutions by tracing the leaders. The leader search refines current optimal solutions by approaching or deviating from the central point of the population and then executes a single-round coordinate descent. The wanderer search carries out comprehensive search space expansion. The crossover search generates offspring using solutions from superior parents. Role learning automates the process in which a search agent decides whether to become a follower or a wanderer. A global optima estimation framework (GOEF) is proposed to offer guidelines for designing an efficient optimization algorithm, and IOA is proved to attain global optima. A differentiable integrated optimization algorithm (DIOA) that extends gradient descent is put forward to train deep learning models. Empirical case studies conclude that IOA shows a much faster convergence speed and finds better solutions than the other 8 comparative algorithms based on 27 benchmark functions. IOA has also been applied to solve unit commitment problems in the power system and shows satisfactory results. A power line sub-image classification model based on a convolutional neural network (CNN) is optimized by DIOA. Compared with the pure gradient descent approach, DIOA converges significantly faster and obtains a high test set accuracy with much fewer training epochs.

Adaptive multi-objective optimization for emergency evacuation at metro stations

Evacuation is a critical issue at metro stations, where damage, or even death, could be caused due to unexpected accidents if without proper evacuation. Multi-objectives are often desired in the evacuation management, and evacuation strategies should be tailored due to the dynamic features of overcrowded passengers and high uncertainty at the metro stations. A simulation-based approach integrating Light Gradient Boosting Machine (LightGBM) and Non-dominated Sorting Genetic Algorithm III (NSGA-III) is proposed to realize the automatic evacuation evaluation and adaptive optimization at metro stations. A framework consisting of 3 objectives and 9 influential factors is developed for the evacuation evaluation and adaptive optimization. A LightGBM based meta-model is used to construct the relationship between influential factors and objectives. A LightGBM and NSGA-III integrated optimization algorithm is employed to automatically evaluate the evacuation events and seek the adaptive strategies for the station renovation in order to achieve a safe evacuation under specific conditions. A station model simulating a metro station in Singapore is constructed to test the effectiveness and applicability of the proposed approach. It is found that (1) Among 50 target station based cases, 33 failing cases (at least one of the objectives fails) are identified, and the automatic evaluation results indicate the studied metro station could successfully evacuate the passengers when the passenger volume is no more than 1000, but it is very likely to fail when a higher volume exists; (2) Adaptive optimization strategies can be found for every different scenario, and an average 24.8% of the improvement degree can be achieved for all scenarios; (3) The adaptive multi-objective optimization is more cost-effective, presenting an average cost efficiency 2.04, which is significantly higher than the average of the non-adaptive optimization cost efficiency, 0.57. The novelty of this research lies in that (a) A LightGBM-based meta-model is built to construct the relationship between desired multi-objectives and the influential factors, which lays the foundation for the evacuation management from multi-objective optimization perspective; (b) The LightGBM and NSGA-III integrated model can realize the automatic evacuation evaluation and accordingly provide the adaptive strategies under specific conditions.

Wind Turbine Bearing Temperature Forecasting Using a New Data-Driven Ensemble Approach

The bearing temperature forecasting provide can provide early detection of the gearbox operating status of wind turbines. To achieve high precision and reliable performance in bearing temperature forecasting, a novel hybrid model is proposed in the paper, which is composed of three phases. Firstly, the variational mode decomposition (VMD) method is employed to decompose raw bearing temperature data into several sub-series with different frequencies. Then, the SAE-GMDH method is utilized as the predictor in the subseries. The stacked autoencoder (SAE) is for the low-latitude features of raw data, while the group method of data handling (GMDH) is applied for the sub-series forecasting. Finally, the imperialist competitive algorithm (ICA) optimizes the weights for subseries and combines them to achieve the final forecasting results. By analytical investigation and comparing the final prediction results in all experiments, it can be summarized that (1) the proposed model has achieved excellent prediction outcome by integrating optimization algorithms with predictors; (2) the experiment results proved that the proposed model outperformed other selective models, with higher accuracies in all datasets, including three state-of-the-art models.

Parameter Identification by High-Resolution Inverse Numerical Model Based on LBM/CMA-ES: Application to Chalk Aquifer (North of France)

The present paper proposes the numerical solution of an inverse problem in groundwater flow (Darcy’s equation). This solution was achieved by combining a high-resolution new code HYSFLO-LBM (Hydrodynamic of Subsurface Flow by Lattice Boltzmann Method), based on LBM, to solve the direct problem, and the metaheuristic optimization algorithm CMA-ES ES (Covariance Matrix Adaptation-Evolution Strategy) to solve the optimization step. The integrated optimization algorithm which resulted from this combination, HYSFLO-LBM/CMA-ES, was applied to the hydrogeological experimental site of Beauvais (Northern France), instrumented by a set of sensors distributed over 20 hydrogeological wells. Hydrogeological parameters measured by the sensors are necessary to understand the aquifer functioning and to serve as input data for the identification of the transmissivity field by the HYSFLO-LBM/CMA-ES code. Results demonstrated an excellent concordance between the integrated optimization algorithm and hydrogeological applied methods (pumping test and magnetic resonance sounding). The spatial distribution of the transmissivity and hydraulic conductivity are related to the heterogeneous distribution of aquifer formations. The LBM and CMA-ES were chosen for their proven excellent performance and lesser cost, in terms of both money and time, unlike the geophysical survey and pumping test. The model can be used and developed as a decision support tool for integrated water resources management in the region.