Search results

Software defect prediction (SDP) is a critical aspect of software quality assurance, aiming to identify and manage potential defects in software systems. In this paper, we have proposed a novel hybrid approach that combines Gray Wolf Optimization with Feature Selection (GWOFS) and multilayer perceptron (MLP) for SDP. The GWOFS-MLP hybrid model is designed to optimize feature selection, ultimately enhancing the accuracy and efficiency of SDP. Gray Wolf Optimization, inspired by the social hierarchy and hunting behavior of gray wolves, is employed to select a subset of relevant features from an extensive pool of potential predictors. This study investigates the key challenges that traditional SDP approaches encounter and proposes promising solutions to overcome time complexity and the curse of the dimensionality reduction problem.

The integration of GWOFS and MLP results in a robust hybrid model that can adapt to diverse software datasets. This feature selection process harnesses the cooperative hunting behavior of wolves, allowing for the exploration of critical feature combinations. The selected features are then fed into an MLP, a powerful artificial neural network (ANN) known for its capability to learn intricate patterns within software metrics. MLP serves as the predictive engine, utilizing the curated feature set to model and classify software defects accurately.

The performance evaluation of the GWOFS-MLP hybrid model on a real-world software defect dataset demonstrates its effectiveness. The model achieves a remarkable training accuracy of 97.69% and a testing accuracy of 97.99%. Additionally, the receiver operating characteristic area under the curve (ROC-AUC) score of 0.89 highlights the model’s ability to discriminate between defective and defect-free software components.

Experimental implementations using machine learning-based techniques with feature reduction are conducted to validate the proposed solutions. The goal is to enhance SDP’s accuracy, relevance and efficiency, ultimately improving software quality assurance processes. The confusion matrix further illustrates the model’s performance, with only a small number of false positives and false negatives.

While the Chinese securities market is booming, the phenomenon of listed companies falling into financial distress is also emerging, which affects the operation and development of enterprises and also jeopardizes the interests of investors. Therefore, it is important to understand how to accurately and reasonably predict the financial distress of enterprises.

In the present study, ensemble feature selection (EFS) and improved stacking were used for financial distress prediction (FDP). Mutual information, analysis of variance (ANOVA), random forest (RF), genetic algorithms, and recursive feature elimination (RFE) were chosen for EFS to select features. Since there may be missing information when feeding the results of the base learner directly into the meta-learner, the features with high importance were fed into the meta-learner together. A screening layer was added to select the meta-learner with better performance. Finally, Optima hyperparameters were used for parameter tuning by the learners.

An empirical study was conducted with a sample of A-share listed companies in China. The F1-score of the model constructed using the features screened by EFS reached 84.55%, representing an improvement of 4.37% compared to the original features. To verify the effectiveness of improved stacking, benchmark model comparison experiments were conducted. Compared to the original stacking model, the accuracy of the improved stacking model was improved by 0.44%, and the F1-score was improved by 0.51%. In addition, the improved stacking model had the highest area under the curve (AUC) value (0.905) among all the compared models.

Compared to previous models, the proposed FDP model has better performance, thus bridging the research gap of feature selection. The present study provides new ideas for stacking improvement research and a reference for subsequent research in this field.

As data centres grow in size and complexity, traditional air-cooling methods are becoming less effective and more expensive. Immersion cooling, where servers are submerged in a dielectric fluid, has emerged as a promising alternative. Ensuring reliable operations in data centre applications requires the development of an effective control framework for immersion cooling systems, which necessitates the prediction of server temperature. While deep learning-based temperature prediction models have shown effectiveness, further enhancement is needed to improve their prediction accuracy. This study aims to develop a temperature prediction model using Long Short-Term Memory (LSTM) Networks based on recursive encoder-decoder architecture.

This paper explores the use of deep learning algorithms to predict the temperature of a heater in a two-phase immersion-cooled system using NOVEC 7100. The performance of recursive-long short-term memory-encoder-decoder (R-LSTM-ED), recursive-convolutional neural network-LSTM (R-CNN-LSTM) and R-LSTM approaches are compared using mean absolute error, root mean square error, mean absolute percentage error and coefficient of determination (R²) as performance metrics. The impact of window size, sampling period and noise within training data on the performance of the model is investigated.

The R-LSTM-ED consistently outperforms the R-LSTM model by 6%, 15.8% and 12.5%, and R-CNN-LSTM model by 4%, 11% and 12.3% in all forecast ranges of 10, 30 and 60 s, respectively, averaged across all the workloads considered in the study. The optimum sampling period based on the study is found to be 2 s and the window size to be 60 s. The performance of the model deteriorates significantly as the noise level reaches 10%.

The proposed models are currently trained on data collected from an experimental setup simulating data centre loads. Future research should seek to extend the applicability of the models by incorporating time series data from immersion-cooled servers.

The proposed multivariate-recursive-prediction models are trained and tested by using real Data Centre workload traces applied to the immersion-cooled system developed in the laboratory.

The purpose of this study is to propose a precise and standardized strategy for numerically simulating vehicle aerodynamics.

Error sources in computational fluid dynamics were analyzed. Additionally, controllable experiential and discretization errors, which significantly influence the calculated results, are expounded upon. Considering the airflow mechanism around a vehicle, the computational efficiency and accuracy of each solution strategy were compared and analyzed through numerous computational cases. Finally, the most suitable numerical strategy, including the turbulence model, simplified vehicle model, calculation domain, boundary conditions, grids and discretization scheme, was identified. Two simplified vehicle models were introduced, and relevant wind tunnel tests were performed to validate the selected strategy.

Errors in vehicle computational aerodynamics mainly stem from the unreasonable simplification of the vehicle model, calculation domain, definite solution conditions, grid strategy and discretization schemes. Using the proposed standardized numerical strategy, the simulated steady and transient aerodynamic characteristics agreed well with the experimental results.

Building upon the modified Low-Reynolds Number k-e model and Scale Adaptive Simulation model, to the best of the authors’ knowledge, a precise and standardized numerical simulation strategy for vehicle aerodynamics is proposed for the first time, which can be integrated into vehicle research and design.

Weak repeatability is observed in handcrafted keypoints, leading to tracking failures in visual simultaneous localization and mapping (SLAM) systems under challenging scenarios such as illumination change, rapid rotation and large angle of view variation. In contrast, learning-based keypoints exhibit higher repetition but entail considerable computational costs. This paper proposes an innovative algorithm for keypoint extraction, aiming to strike an equilibrium between precision and efficiency. This paper aims to attain accurate, robust and versatile visual localization in scenes of formidable complexity.

SiLK-SLAM initially refines the cutting-edge learning-based extractor, SiLK, and introduces an innovative postprocessing algorithm for keypoint homogenization and operational efficiency. Furthermore, SiLK-SLAM devises a reliable relocalization strategy called PCPnP, leveraging progressive and consistent sampling, thereby bolstering its robustness.

Empirical evaluations conducted on TUM, KITTI and EuRoC data sets substantiate SiLK-SLAM’s superior localization accuracy compared to ORB-SLAM3 and other methods. Compared to ORB-SLAM3, SiLK-SLAM demonstrates an enhancement in localization accuracy even by 70.99%, 87.20% and 85.27% across the three data sets. The relocalization experiments demonstrate SiLK-SLAM’s capability in producing precise and repeatable keypoints, showcasing its robustness in challenging environments.

The SiLK-SLAM achieves exceedingly elevated localization accuracy and resilience in formidable scenarios, holding paramount importance in enhancing the autonomy of robots navigating intricate environments. Code is available at https://github.com/Pepper-FlavoredChewingGum/SiLK-SLAM.

In response to the challenge of reduced efficiency or failure of robot motion planning algorithms when faced with end-effector constraints, this study aims to propose a new constraint method to improve the performance of the sampling-based planner.

In this work, a constraint method (TC method) based on the idea of cross-sampling is proposed. This method uses the tangent space in the workspace to approximate the constrained manifold pattern and projects the entire sampling process into the workspace for constraint correction. This method avoids the need for extensive computational work involving multiple iterations of the Jacobi inverse matrix in the configuration space and retains the sampling properties of the sampling-based algorithm.

Simulation results demonstrate that the performance of the planner when using the TC method under the end-effector constraint surpasses that of other methods. Physical experiments further confirm that the TC-Planner does not cause excessive constraint errors that might lead to task failure. Moreover, field tests conducted on robots underscore the effectiveness of the TC-Planner, and its excellent performance, thereby advancing the autonomy of robots in power-line connection tasks.

This paper proposes a new constraint method combined with the rapid-exploring random trees algorithm to generate collision-free trajectories that satisfy the constraints for a high-dimensional robotic system under end-effector constraints. In a series of simulation and experimental tests, the planner using the TC method under end-effector constraints efficiently performs. Tests on a power distribution live-line operation robot also show that the TC method can greatly aid the robot in completing operation tasks with end-effector constraints. This helps robots to perform tasks with complex end-effector constraints such as grinding and welding more efficiently and autonomously.

This paper aims to develop a multi-objective problem for scheduling the operations of trucks entering and exiting cross-docks where the number of unloaded or loaded products by trucks is fuzzy logistic. The first objective function minimizes the maximum time to receive the products. The second objective function minimizes the emission cost of trucks. Finally, the third objective function minimizes the number of trucks assigned to the entrance and exit doors.

Two steps are implemented to validate and modify the proposed model. In the first step, two random numerical examples in small dimensions were solved by GAMS software with min-max objective function as well as genetic algorithms (GA) and particle swarm optimization. In the second step, due to the increasing dimensions of the problem and computational complexity, the problem in question is part of the NP-Hard problem, and therefore multi-objective meta-heuristic algorithms are used along with validation and parameter adjustment.

Therefore, non-dominated sorting genetic algorithm (NSGA-II) and non-dominated ranking genetic algorithm (NRGA) are used to solve 30 random problems in high dimensions. Then, the algorithms were ranked using the TOPSIS method for each problem according to the results obtained from the evaluation criteria. The analysis of the results confirms the applicability of the proposed model and solution methods.

This paper proposes mathematical model of truck scheduling for a real problem, including cross-docks that play an essential role in supply chains, as they could reduce order delivery time, inventory holding costs and shipping costs. To solve the proposed multi-objective mathematical model, as the problem is NP-hard, multi-objective meta-heuristic algorithms are used along with validation and parameter adjustment. Therefore, NSGA-II and NRGA are used to solve 30 random problems in high dimensions.

Ensuring the safety of structures is important. However, when a structure possesses both an implicit performance function and an extremely small failure probability, traditional methods struggle to conduct a reliability analysis. Therefore, this paper proposes a reliability analysis method aimed at enhancing the efficiency of rare event analysis, using the widely recognized Relevant Vector Machine (RVM).

Drawing from the principles of importance sampling (IS), this paper employs Harris Hawks Optimization (HHO) to ascertain the optimal design point. This approach not only guarantees precision but also facilitates the RVM in approximating the limit state surface. When the U learning function, designed for Kriging, is applied to RVM, it results in sample clustering in the design of experiment (DoE). Therefore, this paper proposes a FU learning function, which is more suitable for RVM.

Three numerical examples and two engineering problem demonstrate the effectiveness of the proposed method.

By employing the HHO algorithm, this paper innovatively applies RVM in IS reliability analysis, proposing a novel method termed RVM-HIS. The RVM-HIS demonstrates exceptional computational efficiency, making it eminently suitable for rare events reliability analysis with implicit performance function. Moreover, the computational efficiency of RVM-HIS has been significantly enhanced through the improvement of the U learning function.

This paper aims to propose a simulation-based performance evaluation model for the drone-based delivery of aid items to disaster-affected areas. The objective of the model is to perform analytical studies, evaluate the performance of drone delivery systems for humanitarian logistics and can support the decision-making on the operational design of the system – on where to locate drone take-off points and on assignment and scheduling of delivery tasks to drones.

This simulation model captures the dynamics and variabilities of the drone-based delivery system, including demand rates, location of demand points, time-dependent parameters and possible failures of drones’ operations. An optimization model integrated with the simulation system can update the optimality of drones’ schedules and delivery assignments.

An extensive set of experiments was performed to evaluate alternative strategies to demonstrate the effectiveness for the proposed optimization/simulation system. In the first set of experiments, the authors use the simulation-based evaluation tool for a case study for Central Florida. The goal of this set of experiments is to show how the proposed system can be used for decision-making and decision-support. The second set of experiments presents a series of numerical studies for a set of randomly generated instances.

The goal is to develop a simulation system that can allow one to evaluate performance of drone-based delivery systems, accounting for the uncertainties through simulations of real-life drone delivery flights. The proposed simulation model captures the variations in different system parameters, including interval of updating the system after receiving new information, demand parameters: the demand rate and their spatial distribution (i.e. their locations), service time parameters: travel times, setup and loading times, payload drop-off times and repair times and drone energy level: battery’s energy is impacted and requires battery change/recharging while flying.

We argue that a fundamental issue regarding how to search and how to switch between different cognitive modes lies in the decision rules that influence the dynamics of learning and exploration. We examine the search logics underlying these decision rules and propose conceptual prompts that can be applied mentally or computationally to aid managers’ decision-making.

By applying Multi-Armed Bandit (MAB) modeling to simulate agents’ interaction with dynamic environments, we compared the patterns and performance of selected MAB algorithms under different configurations of environmental conditions.

We develop three conceptual prompts. First, the simple heuristic-based exploration strategy works well in conditions of low environmental variability and few alternatives. Second, an exploration strategy that combines simple and de-biasing heuristics is suitable for most dynamic and complex decision environments. Third, the uncertainty-based exploration strategy is more applicable in the condition of high environmental unpredictability as it can more effectively recognize deviated patterns.

This study contributes to emerging research on using algorithms to develop novel concepts and combining heuristics and algorithmic intelligence in strategic decision-making.

This study offers insights that there are different possibilities for exploration strategies for managers to apply conceptually and that the adaptability of cognitive-distant search may be underestimated in turbulent environments.

Drawing on insights from machine learning and cognitive psychology research, we demonstrate the fitness of different exploration strategies in different dynamic environmental configurations by comparing the different search logics that underlie the three MAB algorithms.