Название: Applications in Reliability and Statistical Computing
Автор: Hoang Pham
Издательство: Springer
Год: 2023
Страниц: 310
Язык: английский
Формат: pdf (true), epub
Размер: 42.6 MB

This book discusses practical applications of reliability and statistical methods and techniques in various disciplines, using Machine Learning, Artificial Intelligence, optimization, and other computation methods. Bringing together research from international experts, each chapter aims to cover both methods and practical aspects on reliability or statistical computations with emphasis on applications.

5G and IoT are set to generate an estimated 1 billion terabytes of data by 2025 and companies continue to search for new techniques and tools that can help them practice data collection effectively in promoting their business. This book explores the era of Big Data through reliability and statistical computing, showcasing how almost all applications in our daily life have experienced a dramatic shift in the past two decades to a truly global industry.

The requirement for Software Reliability Growth Models (SRGMs) has increased exponentially in response to the growing demand for strong and reliable software systems. During the testing phase of the Software Development Life Cycle (SDLC), SRGMs are particularly effective for estimating fault content, minimizing testing expenses, and maximizing software reliability. There has been a lot of research into selecting the best SRGMs for a certain failure dataset and then ranking all the SRGMs against the dataset. In this chapter, we have studied the mentioned problem and the solution to automate it with the developed compact Decision Support System (DSS), which includes all the functionalities and computational analysis of error logs and ensure error-free software to achieve the desired objective. The DSS is developed in Python utilizing several well-known packages such as Numpy, Scipy, Tkinter, and Pandas. To rank SRGMs employed in the DSS, we used Entropy & Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) ranking methodology. The implemented schema provides highly accurate performance indexes for the SRGMs required for efficient ranking, emphasizing the significance of the proposed prototype of DSS in the open literature, being a novel and ingenious development in the domain of software reliability.

Software reliability is defined as a program's ability to perform its functions under specific conditions for a specific time. Software reliability, according to ANSI, is defined as the likelihood of fault-free software functioning in a predefined environment for a specified time. Software quality, usability, functionality, serviceability, performance, maintainability, capability, and documentation all contribute to software reliability. Due to the high convolution of software, obtaining software reliability is difficult. Due to the rapid growth of system space and the feasibility of upgrading the software, system developers tend to increase the level of complexity within layers of software. It is difficult to achieve a satisfactory level of reliability for a highly complex system containing software. The number of users is directly influenced by software reliability. In practice, a reliable piece of software should have backup (redundant) code that runs the exception handling checks. As a result, the space complexity increases, necessitating more storage, which has an impact on the execution speed. However, for the following reasons, reliability still takes precedence over efficiency since compared to reliability, efficiency can easily be solved by utilizing better hardware resources. A user (or organization) would never use a less reliable or unreliable software that would hinder its growth and productivity. Furthermore, for deadline-based applications, an unreliable software might lead to the loss of invaluable information due to system crash. Origin of unreliability is unclear in a software system with distributed faults through its structure.

Including numerous illustrations and worked examples, the book is of interest to researchers, practicing engineers, and postgraduate students in the fields of reliability engineering, statistical computing, and Machine Learning.