Scholarlink

<p>Software maintenance is the trendsetting activity of modifying, fixing and enhancing the software code deliverables once it is turned over to the client. Evolving with large-scale data available in the system, the software system indulges in deriving the software maintenance. Though software testing and maintenance are critical, it is costly and time-consuming operations in the software development lifecycle. Moreover, the conventional approaches struggle to keep pace with the quick evolution of software systems, resulting in inaccurate or incomplete testing, security vulnerabilities, and undetected bugs. The conventional approaches such as existing testing models, code reviews, and manual testing, often struggle to keep pace with the quick updates and changes to software models, resulting in an important increase in maintenance and testing costs, effort, and time. In addition, the traditional models mostly depend on simplistic assumptions about software behaviour, ignoring the complex relationships among the environmental factors, user interactions, and software components. This oversimplification can result in inaccurate or incomplete testing, leading to unidentified bugs, security concerns, and performance problems that can have serious problems in critical software systems. Therefore, there is an urgent requirement for innovative models that can efficiently maintain and test the software systems, guaranteeing their performance, security, and reliability. This work develops a novel Adaptive Ensemble Deep Learning (AEDL) and Statistical Fitness-based Fire Hawk Optimizer (SFFHO) for revolutionizing software testing and maintenance. This work presents a new software testing and maintenance approach utilizing a deep learning approach. At first, the significant data is fetched from available data resources. Further, the collected data is subjected to the Adaptive Ensemble Deep Learning (AEDL) model. Here, the AEDL model is built with the combination of the Variational Autoencoder (VAE), Attention-based Convolutional Neural Network (ACNN) and Dilated Recurrent Neural Network (DRNN). The developed model helps maintain and test the software based on the client's requirements. Moreover, to enhance the system, the hyper-parameter of AEDL is tuned by the Statistical Fitness-based Fire Hawk Optimizer (SFFHO). Finally, the designed system is contrasted with other traditional approaches that prove better efficiency.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

SFFHO: Development of Statistical Fitness-based Fire Hawk Optimizer for Software Testing and Maintenance Approach using Adaptive Deep Learning Method

  • 期刊论文

  • 发表日期：

    2025年6月4日

查看全文
  • G. Prabu,
  • C. Sujatha,
  • J. Erin Shine,
  • T. Arulkumar

摘要

Software maintenance is the trendsetting activity of modifying, fixing and enhancing the software code deliverables once it is turned over to the client. Evolving with large-scale data available in the system, the software system indulges in deriving the software maintenance. Though software testing and maintenance are critical, it is costly and time-consuming operations in the software development lifecycle. Moreover, the conventional approaches struggle to keep pace with the quick evolution of software systems, resulting in inaccurate or incomplete testing, security vulnerabilities, and undetected bugs. The conventional approaches such as existing testing models, code reviews, and manual testing, often struggle to keep pace with the quick updates and changes to software models, resulting in an important increase in maintenance and testing costs, effort, and time. In addition, the traditional models mostly depend on simplistic assumptions about software behaviour, ignoring the complex relationships among the environmental factors, user interactions, and software components. This oversimplification can result in inaccurate or incomplete testing, leading to unidentified bugs, security concerns, and performance problems that can have serious problems in critical software systems. Therefore, there is an urgent requirement for innovative models that can efficiently maintain and test the software systems, guaranteeing their performance, security, and reliability. This work develops a novel Adaptive Ensemble Deep Learning (AEDL) and Statistical Fitness-based Fire Hawk Optimizer (SFFHO) for revolutionizing software testing and maintenance. This work presents a new software testing and maintenance approach utilizing a deep learning approach. At first, the significant data is fetched from available data resources. Further, the collected data is subjected to the Adaptive Ensemble Deep Learning (AEDL) model. Here, the AEDL model is built with the combination of the Variational Autoencoder (VAE), Attention-based Convolutional Neural Network (ACNN) and Dilated Recurrent Neural Network (DRNN). The developed model helps maintain and test the software based on the client's requirements. Moreover, to enhance the system, the hyper-parameter of AEDL is tuned by the Statistical Fitness-based Fire Hawk Optimizer (SFFHO). Finally, the designed system is contrasted with other traditional approaches that prove better efficiency.