xxxv Chapter 2 presents modern computing paradigms of an intelligent system that handles imprecision as well as provides optimized outcomes. The chapt

Transcription

1 xxxiv Preface Meta-Heuristics Optimization (MO) techniques are attractive global optimization methods inspired by the various phenomena arising in nature and man-made problems. They include Fuzzy Logic, Artificial Neural Networks, Firefly, Simulated Annealing, Tabu Search, Variable Neighborhood Search, Support Vector Machine, Genetic Algorithms, Memetic Algorithms, Differential Evolutions, Particle Swarm Optimization, Glowworm Swarm Optimization, Bee Algorithms, Bacterial Foraging, Ant Colony Algorithms, etc. These are relatively newer additions to the class of numerical optimization algorithms. These methods have been successfully applied to a wide range of real-world application problems. Natural disasters and man-made chaotic problems partially can be solved by classical optimization techniques. Modern optimization algorithms are capable of handling and tackling these problems with a higher level of satisfaction. This book brings forward and highlights the cutting-edge developments in this promising research area by bringing together researchers, scientists, decision makers, managers, and practitioners. President Barack Obama has mentioned the imperfect world and Prime Minister David Cameron has mentioned robust economy. Their words of wisdom will be very useful for readers across the globe to face the new and greatest challenges in the future. Meta-Heuristics Optimization methods have been very successful in tackling a variety of optimization problems in areas such as Industry, Business, Logistics, Computer Science, Engineering, Government, etc. For this book project, we have included chapters contributing to methodological developments and successful implementations of Heuristics/ Metaheuristics and their hybrids. Special emphasis is on real world applications problems, but other implementations will also been considered. This book is very useful for graduate and postgraduate students, decision makers and researchers in private sectors, universities, and industries in the field of various sciences and management such as mathematics/applied mathematics, physics, chemistry, computer science, management, business, economics, and finance, wherever one wants to model their uncertain practical and real life problems. It is well known that uncertainty is inevitable in every field of engineering, management, and science. This book will be of significance and important in handling the real world global problems. This book includes 25 excellent peer-reviewed chapters in the research areas of Artificial Intelligent Algorithms and Techniques for Handling Uncertainties. Brief discussions on all the chapters follow: Chapter 1 proposes a robust control approach for the class of chaotic systems subject to magnitude and rate actuator constraints. Numerical simulations are performed to validate the approach, applying it to the Lorenz chaotic system and to a chaotic aeroelastic system, and parameter uncertainties are also considered to prove its robustness. The results confirm the effectiveness of the approach and the constraints are guaranteed as opposed to other control techniques, which do not consider any kind of constraints.

2 xxxv Chapter 2 presents modern computing paradigms of an intelligent system that handles imprecision as well as provides optimized outcomes. The chapter extensively discusses the role of the Genetic Algorithm (GA) in search and optimization processes along with discussion of applications developed so far. The chapter presents the theory of Fuzzy Logic and its role in linguistic knowledge representation. A very detailed discussion on Fuzzy Rule-Based System is presented along with major applications developed in different domains. The chapter presents algorithms of implementing intelligent procedure to decide whether a patient is prone to heart disease or not. The significant advantage of the presented research work is that applications that do not have any mathematical formulation and still demand optimization can be easily solved using the designed approach. Chapter 3 presents a two-phase mathematical programming approach for effective supply chain design with product life cycle uncertainty considerations. Supply chain is an alliance of independent business processes, such as supplier, manufacturing, and distribution processes, that perform the critical functions in the order fulfillment process. However, the discussions in marketing and logistic literature universally conclude that it would be desirable to determine the life cycle of products in the firm, as they have a great impact on appropriate supply chain design. Designing a supply chain effectively is a complex and challenging task, due to the increasing outsourcing, globalization of businesses, continuous advances in information technology, and product life cycle uncertainty. Indeed, uncertainty is one of the characteristics of the product life cycle. In particular, the strategic design of the supply chain has to take uncertain information into account. The aim of Chapter 4 is to describe a swarm-based optimization algorithm called the Bees Algorithm (BA) and its applications on real world problems. After an explanation of the natural foraging behavior of honeybees, the basic Bees Algorithm and its enhanced version based on Adaptive Neighborhood Search and Site Abandonment (ANSSA) strategy are described and two applications are discussed in detail. The first application deals with the optimization of several benchmark functions and the results obtained by the ANSSA-based BA are compared with the basic BA and other optimization algorithms. The second application deals with the multi-objective optimization problem in finding the best supply chain configuration. The offer of innovative technologies and the growth of demands to new services, especially those with higher transmission rates, make the access network planning an important stage in the evolution of cellular systems. Several technological options of transmission systems are already available and to choose the best among them is a great challenge for network planners. Chapter 5 presents a study for strategic planning of the interconnection of base stations in a cellular mobile network. Chapter 6: Clinical Decision Support Systems (CDSS) are widely applied in healthcare processes, raising several challenges related not only to sensor technology, hardware, software, and communications, but also to the management of the collected, processed, and presented data. Thus, in this chapter, machine-learning techniques are applied into CDSS aiming to establish knowledge refinement and discovery with the purpose of giving reliable explanations and support to healthcare professionals and patients. In line with this, there are promising methodologies with capabilities to deal with missing or noise data, as well as with the ability to produce reliable explanations based on a small amount of data. Meta-Heuristics Optimization (MO) techniques are attractive global optimization methods inspired by the various industrial phenomena with uncertainty. These methods have been successfully applied to a wide range of chemical engineering problems with a higher level of satisfaction. In Chapter 7, the authors introduce multiple artificial intelligence techniques: Genetic Algorithm (GA), Biogeography- Based Optimization (BBO), Differential Evolution (DE), Evolutionary Strategy (ES), Probability-Based

3 xxxvi Incremental Learning (PBIL), Stud Genetic Algorithm (SGA), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), Artificial Bee Colony (ABC), and Fuzzy Logic (FL). The chapter includes the introduction of algorithms and their applications to handle uncertainty in the chemical process operation. A robust fuzzy digital PID control methodology based on gain and phase margins specifications is proposed. A mathematical formulation based on gain and phase margins specifications, the Takagi- Sugeno fuzzy model of the process to be controlled, the structure of the digital PID controller, and the time delay uncertain system are developed. A multiobjective genetic strategy is defined to tune the fuzzy digital PID controller parameters, so the gain and phase margins specified to the fuzzy control system are found. An analysis of necessary and sufficient conditions for fuzzy digital PID controller design with robust stability, with the proposal of the two theorems, is presented in Chapter 8. Chapter 9: The direct focus of this chapter is to explore the potential of online Adaptive NeuroFuzzy Type-2 (ANFT2) control system for damping inter-area oscillations using Static Synchronous Compensator (STATCOM). The performance evaluation of proposed control strategy has been validated using two and three machine power systems. The nonlinear time domain simulations reveal that ANFT2 has excellent damping capabilities as compared to conventional PI control. Simulation results for different performance indices further emphasize the optimal performance of ANFT2 with uncertain mean and variance of triangular membership function in transient and steady state region. Chapter 10: In this chapter, the author presents a new neural stereo matching method using very high resolution IKONOS images. A neural field is chosen due to its good management of imprecision and uncertainty relatives to real problems in general and to this one in particular. To show the effectiveness of a proposed method, the chapter contains at first details about encountered problems, and secondly, it explains the stereo matching process, its different kinds, and a chosen approach; thirdly, it gives obtained results using panchromatic and colour images. The aim of this Chapter 11 is to discuss the application of fuzzy logic for mapping the Agro-Ecological Zones (AEZ). Agricultural land is one of the most important parts in agriculture because it determines the type of suitable plants for that land as well as its productivity. Fuzzy logic has been applied to solve various problems and fields including agricultural area. In this study, fuzzy logic is implemented to map AEZ in Boyolali. In Chapter 12, the authors propose a new scheme of fusing the weak Priestley-Chao Kernel Estimators (PCKEs) based on Choquet fuzzy integral, which differs from all the existing models of regressor fusion. The new scheme uses Choquet fuzzy integral to fuse several target outputs from different PCKEs, in which the optimal bandwidths are obtained with cross-validation criteria. The key of applying fuzzy integral to PCKE fusion is the determination of fuzzy measure. Considering the advantage of Particle Swarm Optimization (PSO) algorithm on convergence rate, the authors use three different PSO algorithms (i.e., Standard PSO [SPSO], Gaussian PSO [GPSO], and GPSO with Gaussian Jump [GPSOGJ]) to determine the general and λ fuzzy measures. The experimental results on the standard testing functions and practical Fourier Transform Infrared Spectroscopy (FTIR) datasets show that the new paradigm for regression ensemble based on fuzzy integral is more accurate and stable in comparison with any individual PCKE and the Basic Ensemble Method (BEM). This demonstrates the feasibility and effectiveness of the proposed regression ensemble model. Motivated by the fact that stocks are sometimes in great demand but short supply at depots, the Multi- Depot Vehicle Routing Problem with Limited Stocks (MDVRPLS) is introduced in Chapter 13. An effective memetic algorithm and a classical genetic algorithm are developed for solving the MDVRPLS.

4 xxxvii The computational experiments demonstrate that the proposed memetic algorithm can not only produce high-quality solutions within a reasonable computational time but also appear superior to the classical genetic algorithm in terms of solution quality. In Chapter 14, several soft computing techniques are briefly introduced in the research areas of renewable energy and energy efficiency. Then the methodology framework and implementation procedures are presented to demonstrate the application of artificial neural networks and curve fitting for renewable energy network design and optimization, which has the capability to handle restoration during the extreme and emergency situations with the uncertain parameters. In Chapter 15, the author has considered the problem of sequence-dependent setup times hybrid flowshop scheduling with the objectives of minimizing the makespan and the sum of earliness/tardiness of jobs, and developed a multi-objective Hybrid Metaheuristic (HMH). This approach is combined with a weighting method that helps to generate many solutions on the non-dominated front. The effectiveness of HMH is tested on 252 benchmark problems. The computational results show that HMH performs better than SALS presented previously. In addition, the possibility of configuring the hybrid metaheuristic behavior provides a certain level of flexibility to adapt the algorithm to the specific characteristics of the instances to be solved. A hybrid unsupervised learning algorithm, which is termed as Evolutionary Rough Multi-Objective Optimization (ERMOO) algorithm, is proposed in Chapter 16. It comprises a judicious integration of the principles of the rough sets theory with the archived multi-objective simulated annealing approach. A measure of the amount of domination between two solutions is incorporated in this chapter to determine the acceptance probability of a new solution with an improvement in the spread of the non-dominated solutions in the Pareto-front by adopting rough sets theory. The performance is demonstrated on reallife breast cancer dataset for identification of Cancer Associated Fibroblasts (CAFs) within the tumor stroma, and the identified biomarkers are reported. Multi-objective optimization is one of the most popular research areas in the world of manufacturing. It concerns the manufacturing optimization problems involving more than one optimization simultaneously. Therefore, to tackle this problem, Chapter 17 proposes a new integrated approach by combining Standard Deviation Method with Particle Swarm Optimization. Two examples of optimizing the advanced manufacturing process parameters are performed to test the proposed approach. The examples considered for this approach are also attempted using other established optimization techniques such as Desirability-based RSM and SDM-GA. The results verify the effectiveness of the proposed approach during multi-objective manufacturing process parameter optimization. Chapter 18 describes and uses different meta-heuristics optimization methods to solve the redundancy optimization problem for multi-state series-parallel power systems. The authors have considered the case where redundant power components are chosen to achieve a desirable level of reliability. The power components of the system are characterized by their cost, capacity, and reliability. The new approach has the advantage of allowing power components with different parameters to be allocated in power systems. To allow fast reliability estimation, a Moment Generating Function (MGF) method is applied. In the past two decades, Swarm Intelligence (SI)-based optimization techniques have drawn the attention of many researchers for finding an efficient solution to optimization problems. Swarm intelligence techniques are characterized by their decentralized way of working that mimics the behavior of colony of ants, swarm of bees, flock of birds, or school of fishes. Algorithmic simplicity and effectiveness of swarm intelligence techniques have made it a powerful tool for solving global optimization problems. Simulation studies of the graceful, but unpredictable, choreography of bird flocks led to the design of

5 xxxviii the particle swarm optimization algorithm. Studies of the foraging behavior of ants resulted in the development of ant colony optimization algorithm. Chapter 19 provides insight into swarm intelligence techniques, specifically particle swarm optimization and its variants. The objective of this chapter is twofold: First, it describes how swarm intelligence techniques are employed to solve various optimization problems. Second, it describes how swarm intelligence techniques are efficiently applied for clustering, by imposing clustering as an optimization problem. The main objective of Chapter 20 is to present a novel hybrid GA-GSA algorithm to permit the reliability analysts or system managers to increase the performance of the system by utilizing uncertain and imprecise data. As most of the data are collected from the historical records or logbooks and are out of date, the analysis conducted based on that contains a lot of uncertainties. Thus, the corresponding results obtained do not tell the exact nature of the system. Therefore, to handle this issue, the proposed algorithm maximizes the Reliability, Availability, and Maintainability (RAM) parameters simultaneously for increasing the performance and productivity of the system. To illustrate the methodology, a numerical example has been applied for the pulping unit of the paper industry, a complex repairable industrial system, situated in the northern part of India. Vehicle routing is a difficult combinatorial optimization problem that has attracted many researchers to apply meta-heuristics to find approximate solutions. In the case of time windows within which goods have to picked up or delivered, it is even more difficult to find good solutions. Of course, congestion makes it hard for planners to find good routes for delivery or pick up because travel times between customers or between a depot and a customer are uncertain. In this chapter, the problem is handled by assigning a range of possible travel times between customers to represent the uncertainty. This social type of routing is hardly studied in literature but makes up the topic of Chapter 21. The 2011 flooding in Thailand was the major inspiration for this research work. Chapter 22 is a descriptive study that examines and explains the common fuzzy logic applications in the medical field after an introduction to fuzzy logic. The medical decision-making process is fuzzy in its nature. The physician handles linguistic concepts in deciding the diagnosis and prognosis. The conversion from this fuzzy nature into crisp real world outcome causes the loss of precision. Fuzzy logic is a suitable way to provide the physician with the support he needs in handling linguistic concepts and get rid of the loss of precision. Fuzzy logic technologies are applied to each area of medicine, and they have been proven to be successful. Chapter 23 presents a new model for baseline detection in cursive languages (Arabic, Kurdish, Persian, etc.) for handwritten Optical Character Recognition (OCR). Several methods exist for baseline detection, such as baseline detection methods based on horizontal projection, based on word skeleton, based on word contour representation, and based on Principal Components Analysis (PCA). Usually, baseline detection is an important part of the preprocessing stage in OCR systems. The baseline detection method, which is discussed in this chapter, uses a heuristic solution and has been tested on handwritten words. Chapter 24 examines the capability of RVM and GPR for prediction of r and e. RVM and GPR models have been developed by using 32 datasets. Radial basis function has been adopted for developing the RVM and GPR models. The developed RVM and GPR give excellent performance. The user can use the developed equation for practical purpose. The developed RVM uses less tuning parameters compared to the GPR. Prediction uncertainty can be measured from the obtained variance. Sensitivity analysis shows that v is the most important parameter for prediction of r and e. This study shows that the developed RVM and GPR can be used to solve different problems in engineering.

6 xxxix Chapter 25 has a novel concept for early warning system design within company, applicable in different industries. The core of the proposed framework is a hybrid fuzzy expert system, which can contain a variety of data mining predictive models responsible for some specific areas and additions to traditional rule blocks. As part of the description of how early warning systems really work, structured techniques and Social Network Analysis (SNA) were introduced with appliance examples pointing to selected software tools for a fast start. The concept of methodology for creating EWS, selecting key indicators, using fuzzy expert systems, and how to improve systems using SNA metrics are explained and discussed with future research directions. This book project offers a golden opportunity for the researchers in the field of Artificial Intelligent Algorithms and Techniques for Handling Uncertainties to disseminate their original, quality, novel, and innovative research findings and results to the global researchers. The editor of this book sincerely thanks all the contributors for their marvelous and invaluable and knowledgeable contributions in making this book a global reference for engineers, scientist, practitioners, economists, financiers, researchers, decision makers, implementers, technologies, and managers in the research areas of Artificial Intelligent Algorithms and Techniques for Handling Uncertainties. Pandian Vasant Universiti Teknologi Petronas, Malaysia