Preface

With the design of every new product, the world is witnessing the continuous development brought on by cross-disciplinary technologies. Instead of taking raw materials and sending them through a real manufacturing process that repeatedly combats tolerances, errors, and energy consumption to arrive at the final product, the assembly details can be directly input into the computation model in order to obtain the material characteristics as output to reduce effort and process costs. To ensure maximum reliability of product development, it is desired that the manufacturing process be driven by optimization. However, even though optimization has previously been applied for various fields, over the past two decades, computational optimization has become very popular for industrial optimizations. Computational intelligence-based optimization is one of several computational techniques that help achieve sustainability in product design and development phases. Among computational intelligence-based techniques, metaheuristic optimization is found to be specifically suitable for industrial optimizations. There are mainly two types of metaheuristic approaches; single-solution based and population based. As per the applications in the field of industrial optimization, this book mainly focuses on population-based (swarm intelligence) metaheuristic approaches.

Swarm intelligence is an important sub-area of optimization that helps develop sustainable materials at nano-, micro-, meso- and macro-levels by identifying the optimum values for different parameters. With the exponential rise in demand for sustainable materials for various purposes, optimization has played an important role over the last few years. Not only is materials data available for researchers and scientists, but sufficient processing resources are also available, which need to be optimized through AI techniques.

Traditional techniques employed by researchers are often cumbersome, expensive and lack sustainability. Hence, there is always a need for having recourse to time-efficient, fail-safe, cheaper intelligent technologies to address problems and ensure long-term sustainability. Since the existing literature available in this respect is nonexistent, this book is proposed to serve as a treatise and knowledge base for the community to inspire them to adapt environment-friendly and sustainable solutions for the future.

This book focuses on developing advanced computational intelligence algorithms for the analysis of data involved in reliability engineering, material design, and manufacturing with the objective of ensuring sustainability. It reveals applications of different models of evolutionary algorithms in the field of optimization with the objective of solving problems to help the manufacturing industries. Some special features of this book include a comprehensive guide for utilizing computational models for reliability engineering, state-of-the-art swarm intelligence methods for solving manufacturing processes and developing sustainable materials, high-quality and innovative research contributions, and a guide for applying computational optimization to reliability and maintainability theory. A chapter-wise summary of the information presented herein follows.

Chapter 1 presents a stochastic model for reliability indices of a computer system with priority and server failure. The model is analyzed by using the semi-Markov process and regenerative point technique. The reliability indices, such as mean time to system failure (MTCSF), availability, busy period of the server due to hardware repair and software upgradation, expected number of treatments given to the server, expected number of hardware repair, and software upgradation, are obtained for arbitrary values of the parameters. The profit analysis of the system model has also been carried out to discern the usefulness of the system under different parametric situations.

Chapter 2 presents a study that optimizes the availability of a turbine unit (TU) of a steam turbine power plant (STPP) using mathematical modeling and a genetic algorithm. The mathematical model is developed using the Markovian birth-death process (MBDP) and Chapman-Kolmogorov differential equations derived for the proposed model. The analytical solution of the mathematical model is derived for a particular case by considering exponential distribution for random variables associated with failure and repair rates. By using a nature-inspired algorithm (NIA), namely a genetic algorithm (GA), an effort is made to attain the global solution of the TU.

Chapter 3 covers the development of the Laplacian artificial bee colony (LABC) algorithm for effective harmonic estimator design. For designing the estimator, a hybrid approach based on least square error minimization with the help of a new version of the artificial bee colony algorithm is proposed. The proposed version employs a Laplacian factor-based update equation in the scout bee phase. For proving the modification meaningful, first the proposed algorithm is tested on several standard benchmark problems, and then it is applied to the estimator design problem. Results reported in on both parts indicate that the proposed modification is meaningful and the performance of the LABC algorithm is comparable with that of many other state-of-the-art algorithms.

Chapter 4 discusses the applications of the cuckoo search algorithm in reliability optimization, which is a novel nature-inspired algorithm that is used to solve complex optimization problems. The algorithm depends on the brood-parasitic strategy of cuckoo species. The usage of Lévy flights is used to produce new candidate resolutions. It can improve the relationship between exploration and exploitation towards the potential of searching. It can also be used in solving engineering problems such as embedded systems, distribution of networks, and scheduling problems. In this chapter, a study of the reliability of the software at static and runtime is performed and the results are also discussed.

Chapter 5 carries out a performance evaluation of the series-parallel computer system with a Gumbel-Hougaard copula family. To analyze the reliability of the system, the partial differential equations are derived from the system's schematic diagram in which reliability measures of system strength, such as reliability, availability, mean time to failure (MTTF), and cost function, are computed. The MTTF of devices, such as workstation, hub, and router, obeys exponential distribution whereas the corresponding repair time follows two different distributions, namely general and copula distribution. The findings of the study are depicted with the help of suitable diagrams and tabular representations.

Chapter 6 covers the applications of artificial intelligence (AI) in sustainable energy development and utilization. To combat the energy and environmental crises, clean and renewable fuels like biofuels are popular as petrodiesel replacement fuels. Biofuels can be obtained from different feedstocks and are successfully tested in diesel engines. However, several parameters influence the output results during their production and engine testing. The accurate prediction of end results is considered challenging with the traditional techniques. Therefore, AI techniques have emerged as being the most successful in solving nonlinear problems and achieving a high success rate in prediction. In this chapter, different AI techniques that have been successfully used in finding a feasible solution for complex problems in biodiesel production and engine testing are discussed in detail.

Chapter 7 introduces a new joint reliability achievement worth (JRAW), joint reliability reduction worth (JRRW), and joint reliability Fussell-Vesely (JRFV) measure for three multistate components of a multistate system. This is a new approach to detect the joint effect of a group of components in improving system reliability. The differencing technique is used in the proposed measures. A steady-state performance level distribution restricted to the component's states is used to evaluate the proposed measures. The universal generating function (UGF) technique is applied for the evaluation of proposed joint importance measures with suitable examples. Chapter 8 presents some inferences about inverse Rayleigh distribution based on joint progressive Type-II censoring. The maximum likelihood estimation and the corresponding asymptotic confidence interval estimation are used as the classical estimation methods. The Bayes estimates are calculated under the squared error loss function (SELF) using Tierney-Kadane's approximation and Metropolis-Hastings algorithm, along with the construction of Bayes estimates highest posterior density credible intervals. A Markov chain Monte Carlo simulation study is carried out to compare different estimation methods and a real-life problem is discussed for illustrative purposes.

Chapter 9 deals with component reliability estimation through competing risk analysis of fuzzy lifetime data. In many cases, the lifetimes of systems are not precisely observed, or they are reported in "vague" terms. This imprecision or vagueness in data can be dealt with more accurately by incorporating fuzzy concepts. In this chapter, a competing risk analysis of lifetime data is performed by considering lifetimes as fuzzy numbers. Using different membership functions, the authors provide procedures for maximum likelihood and a Bayesian estimation of component reliability. They also evaluate bootstrap confidence intervals and the highest posterior density intervals. To observe the impact of various membership functions on the considered...