Search results

This paper attempts to combine the application of artificial intelligence in predicting and evaluating the classification of surrounding rock grades and provides guidance for subsequent support design and reinforcement support operations.

This paper discusses the use of BPNN as the primary tool, combined with three swarm bionic optimization algorithms (GA, PSO, GWO), to solve stability evaluation and grade prediction of surrounding rock in ultra-deep roadway excavation.

Taking the Great Wall ore group as the core and the Shanghaimiao mining area as the extension, the optimal model is applied to the classification of surrounding rock grade in ultra-deep roadway engineering. Prediction results show that the performance of BPNN models is excellent.

Due to the limitations of geological conditions and construction environment in deep coal mines, the period of roadway excavation is too long, resulting in less data collection.

The prediction results can provide guidance for the excavation method, support scheme correction and reinforcement support scheme design of deep coal mine roadway engineering.

It provides guidance for deep mining of coal mine (the premise of surrounding rock support stability), so as to ensure the economic and safety benefits of coal enterprises.

The neural network is applied to rock mechanics in a deep site for the first time, which is used to solve the prediction direction of surrounding rock grade evaluation. The index of the input layer is determined by combining the “three high and one disturbance” with the on-site construction situation, which is closer to the actual project. The swarm intelligent bionic algorithms are selected to optimize the hyperparameters of back propagation neural network, so as to improve the accuracy of the models. The classification and evaluation system of surrounding rock for the Great Wall ore group is constructed, which is the core of Shanghaimiao mining area in the northwest of China, guiding the dynamic adjustment of on-site excavation and support operations.

Many researchers and analysts are interested in evaluating the performance of a system with a network structure as a decision-making unit. In this regard, fuzzy network data envelopment analysis (FNDEA) is a noticeable and worthy method for evaluating the efficiency of a system with fuzzy data. Based on the structure of a fuzzy network system, which consists of at least two serial stages, an intermediate factor has an output nature for the first stage and an input nature for the second stage. Hence, it is inappropriate to allocate the same weight for each stage using this factor. Unfortunately, contrary to real-world conditions, all previous conventional FNDEA studies have considered the same role for intermediate factors to linearize or simplify models. For the first time, this study attempts to determine the upper and lower bounds of the overall efficiencies of a fuzzy two-stage series system and its subprocesses with unequal intermediate product weights.

The proposed model remains in its original nature as a complex combinatorial problem in the nonlinear programming category of NP-hard problems. A genetic algorithm (GA) is utilized as a metaheuristic algorithm, and a novel hybrid GA-FNDEA algorithm is presented to solve the problem.

The findings of the study outlined several theoretical contributions and practical implications, including as compensatory property of DEA, determining upper and lower bounds, improving efficiency in nonlinear systems, reducing computational burden, enhancing evolutionary algorithms and retaining real-world conditions.

Contrary to real-world conditions, all previous conventional FNDEA studies have considered the same role for intermediate factors to linearize or simplify models. For the first time, this study attempts to determine the upper and lower bounds of the overall efficiencies of a fuzzy two-stage series system and its subprocesses with unequal intermediate product weights.

The purpose of this study is to investigate resource-constrained multi-project scheduling problems (RCMPSP) involving uncertainty in the form of time-dependent renewable resource reliability. A key focus is to minimize the makespan (completion time) of projects when resources can become unavailable or fail over time at non-constant rates. Accounting for realistic resource reliability seeks to provide scheduling solutions that better reflect potential delays in practical multi-project environments.

A new discrete-time binary integer programming formulation of RCMPSP is expanded to include time-dependent resource reliability and simultaneously evaluate the time-dependent failure rate and constant repair rate of a resource. A new hybrid immune genetic algorithm with local search (HIGALS) is developed to solve this NP-hard problem. HIGALS incorporates a new coding mechanism, initialization method and local search operator.

A case study tests the proposed HIGALS approach. The validity of the mathematical model is confirmed by solving small-sized problems with GAMS software. The proposed HIGALS algorithm is validated by solving small-sized problems and comparing its solutions with GAMS. The superiority of HIGALS is demonstrated by comparing its solutions with six basic algorithms on medium- and large-sized problems. Results show that HIGALS outperforms existing algorithms, achieving an average reduction in makespan of over 11.79%, while maintaining the advantages of genetic, immune and local search algorithms and avoiding their disadvantages.

Considering time-dependent resource reliability can help project managers plan for disruptions and delays in resource-critical projects. HIGALS provides decision support for robust multi-project scheduling.

This study contributes to the field by investigating RCMPSP with time-dependent renewable resource reliability, which reflects real-world uncertainty more accurately. HIGALS presents a novel approach to balance intensification and diversification for this challenging problem.

On account of the flexibility, large working space and system openness, manipulators are often adopted in automatic grinding and polishing operations. In the flexible roboticized polishing process for complex surfaces with narrow spatial structures, such as aero-engine blades, the contact mode between the tool and the workpiece changes with the transformation of the manipulator’s end posture and the alternation of the workpiece curvature, which often leads to processing contact faults. These faults result in the obsolescence of expensive aerospace components and reduced efficiency. The purpose of this study is to collect vibration signals during the machining process and extract fault characteristic parameters for monitoring and diagnosis for diagnosing faults in automated flexible polishing to protect the workpiece.

This paper proposes a whale optimization algorithm (WAO)-support vector machine model based on the support vector machine and WAO. From the original grinding and polishing vibration signal, 11 time-domain features that can reflect the fluctuation of the vibration signal are extracted as detection features.

Experimental results indicate that this method effectively reflects the relationship between contact faults and diagnostic results, demonstrating good real-time performance and diagnostic capability.

This method provides a crucial theoretical basis for real-time fault diagnosis and monitoring in automatic flexible machining, ensuring reliable automatic flexible polishing processes.

This research aims to explore whether or how organizations adopt people analytics (PA), its value and potential socio-technical factors that can enable or hinder PA projects by disrupting and reshaping human resource management. We do this by focusing on the Italian context.

We conduct a scoping review of data collected between 2018 and 2022 via Google Alerts (GA), a content change detection and notification service that is gaining popularity in scholarly research.

Our findings suggest that the diffusion of PA applications in Italy, especially those of a descriptive nature, is growing. Most of the existing PA applications are positioned in a positive technocratic light, envisioning the value of PA for both employees and organizations. The value for the latter appears to be direct, while the value for employees is realized through organizational initiatives. The findings also suggest that while enablers can vary between PA application types, the barriers, especially technological and environmental, are generic for both descriptive and predictive/prescriptive PA applications.

Theoretically, we propose a framework for analyzing PA applications, their values, enablers and barriers. Methodologically, we present and describe in detail a novel approach, drawing on GA that can be used to study PA in specific contexts. Practically, our study serves as a helpful point of reference for managers planning or implementing PA in Italy, for benchmarking PA in Italy over time and for comparative international studies.

This paper aims to propose a lightweight, high-accuracy object detection model designed to enhance seam tracking quality under strong arcs and splashes condition. Simultaneously, the model aims to reduce computational costs.

The lightweight model is constructed based on Single Shot Multibox Detector (SSD). First, a neural architecture search method based on meta-learning and genetic algorithm is introduced to optimize pruning strategy, reducing human intervention and improving efficiency. Additionally, the Alternating Direction Method of Multipliers (ADMM) is used to perform structural pruning on SSD, effectively compressing the model with minimal loss of accuracy.

Compared to state-of-the-art models, this method better balances feature extraction accuracy and inference speed. Furthermore, seam tracking experiments on this welding robot experimental platform demonstrate that the proposed method exhibits excellent accuracy and robustness in practical applications.

This paper presents an innovative approach that combines ADMM structural pruning and meta-learning-based neural architecture search to significantly enhance the efficiency and performance of the SSD network. This method reduces computational cost while ensuring high detection accuracy, providing a reliable solution for welding robot laser vision systems in practical applications.

Cracks are prevalent signs of pavement distress found on highways globally. The use of artificial intelligence (AI) and deep learning (DL) for crack detection is increasingly considered as an optimal solution. Consequently, this paper introduces a novel, fully connected, optimised convolutional neural network (CNN) model using feature selection algorithms for the purpose of detecting cracks in highway pavements.

To enhance the accuracy of the CNN model for crack detection, the authors employed a fully connected deep learning layers CNN model along with several optimisation techniques. Specifically, three optimisation algorithms, namely adaptive moment estimation (ADAM), stochastic gradient descent with momentum (SGDM), and RMSProp, were utilised to fine-tune the CNN model and enhance its overall performance. Subsequently, the authors implemented eight feature selection algorithms to further improve the accuracy of the optimised CNN model. These feature selection techniques were thoughtfully selected and systematically applied to identify the most relevant features contributing to crack detection in the given dataset. Finally, the authors subjected the proposed model to testing against seven pre-trained models.

The study's results show that the accuracy of the three optimisers (ADAM, SGDM, and RMSProp) with the five deep learning layers model is 97.4%, 98.2%, and 96.09%, respectively. Following this, eight feature selection algorithms were applied to the five deep learning layers to enhance accuracy, with particle swarm optimisation (PSO) achieving the highest F-score at 98.72. The model was then compared with other pre-trained models and exhibited the highest performance.

With an achieved precision of 98.19% and F-score of 98.72% using PSO, the developed model is highly accurate and effective in detecting and evaluating the condition of cracks in pavements. As a result, the model has the potential to significantly reduce the effort required for crack detection and evaluation.

The proposed method for enhancing CNN model accuracy in crack detection stands out for its unique combination of optimisation algorithms (ADAM, SGDM, and RMSProp) with systematic application of multiple feature selection techniques to identify relevant crack detection features and comparing results with existing pre-trained models.

For over 25 years, the United States Green Building Council (USGBC) has significantly influenced the US sustainable construction through its leadership in energy and environmental design (LEED) certification program. This study aims to delve into how Baton Rouge, Louisiana, fares in green building adoption relative to other US capital cities and regions.

The study leverages statistical and geospatial analyses of data sourced from the USGBC, among other databases. It scrutinizes Baton Rouge’s LEED criteria performance using the mean percent weighted criteria to pinpoint the LEED criteria most readily achieved. Moreover, unique metrics, such as the certified green building per capita (CGBC), were formulated to facilitate a comparative analysis of green building adoption across various regions.

Baton Rouge’s CGBC stands at 0.31% (C+), markedly trailing behind the frontrunner, Santa Fe, New Mexico, leading at 3.89% (A+) and in LEED building per capita too. Despite the notable concentration of certified green buildings (CGBs) within Baton Rouge, the city’s green building development appears to be in its infancy. Innovation and design was identified as the most attainable LEED benchmark in Baton Rouge. Additionally, socioeconomic factors, including education and income per capita, were associated with a mild to moderate positive correlation (0.25 = r = 0.36) with the adoption of green building practices across the capitals, while sociocultural infrastructure exhibited a strong positive correlation (r = 0.99).

This study is beneficial to policymakers, urban planners and developers for sustainable urban development and a reference point for subsequent postoccupancy evaluations of CGBs in Baton Rouge and beyond.

This study pioneers the comprehensive analysis of green building adoption rates and probable influencing factors in capital cities in the contiguous US using distinct metrics.

The field of medical technology is constantly evolving, leading to improvements in implantation techniques that offer innovative solutions for treating bone tissue defects. The purpose of this study is to investigate the integration of nano-silica into ceramic scaffolds to enhance the mechanical strength of Hydroxyapatite structures.

Using the design of experiment methodology, 13 distinct ceramic pastes, each optimized for specific mechanical characteristics, are formulated. Rheological testing is performed to ensure suitability for 3D printing, and the pastes are evaluated using techniques such as scanning electron microscopy and energy dispersive X-ray spectroscopy. The definitive screening design optimizer is used to determine an ideal material combination based on factors like extrudability, printability, strength and biocompatibility.

Scaffolds with the optimized HA/SiO₂ composition are fabricated and tested for compression strength, achieving 7.8 MPa.

The research endeavors detailed within this study represent a notable advancement in the augmentation of ceramic scaffold properties tailored for bone tissue engineering applications, particularly focusing on their suitability for integration within load-bearing structures. A particular emphasis is placed on the enhancement of printability, thereby facilitating streamlined fabrication processes.

Predicting commodity futures trading volumes represents an important matter to policymakers and a wide spectrum of market participants. The purpose of this study is to concentrate on the energy sector and explore the trading volume prediction issue for the thermal coal futures traded in Zhengzhou Commodity Exchange in China with daily data spanning January 2016–December 2020.

The nonlinear autoregressive neural network is adopted for this purpose and prediction performance is examined based upon a variety of settings over algorithms for model estimations, numbers of hidden neurons and delays and ratios for splitting the trading volume series into training, validation and testing phases.

A relatively simple model setting is arrived at that leads to predictions of good accuracy and stabilities and maintains small prediction errors up to the 99.273th quantile of the observed trading volume.

The results could, on one hand, serve as standalone technical trading volume predictions. They could, on the other hand, be combined with different (fundamental) prediction results for forming perspectives of trading trends and carrying out policy analysis.