Ebook: Integrated Intelligent Systems for Engineering Design

Background

With the ever increasing complexity of products and customer demands, companies are adopting new strategies to meet the changing technological requirements, shorter product life cycles, and globalization of manufacturing operations. Product design requires more sophisticated procedures and processes and requires designers and engineers possessing different expertise, knowledge and experience to work together. To address these challenges, techniques based on artificial intelligence (AI) are increasingly being used to improve effectiveness and efficiency in the product-design life cycle. Intelligent systems can be beneficially applied to many aspects of design and also design-related tasks at different stages; for example, identifying customer demands and requirements, design and planning, production, delivery, marketing and customer services, etc. Individual intelligent paradigms (such as fuzzy logic, neural network, genetic algorithm, case-based reasoning, and especially expert systems) have been applied to specific stages of the design process (product planning, conceptual design, detailed design). However, increasingly, hybrid solutions that integrate multiple individual intelligent techniques are required to solve complex design problems. The integrated intelligent environment can provide various types of information and knowledge for supporting rapid and intelligent decision-making throughout the entire design process. This is in line with the evolutionary trends of the product design process, from the traditional CAD systems into the knowledge-based engineering and integrated intelligent design systems through a combination of concurrent engineering, collaborative engineering and integrated intelligent techniques.

In recent years, with the advancement of artificial intelligence and information science and technology, there has been a resurgence of work in combining individual intelligent paradigms (knowledge-based systems, fuzzy logic, neural networks, genetic algorithms, case-based reasoning, machine learning and knowledge discovery, data mining algorithms, intelligent agents, soft computing, user intelligent interfaces, etc.) into integrated intelligent systems to solve complex problems. Hybridization of different intelligent systems is an innovative approach to constructing computationally intelligent systems consisting of artificial neural networks, fuzzy inference systems, approximate reasoning and derivative-free optimization methods such as evolutionary computation and so on. The integration of different learning and adaptation techniques, to overcome individual limitations and achieve synergetic effects through hybridization or fusion of these techniques, has contributed to a large number of new intelligent system designs. Hybrid intelligent systems are becoming a very important problem-solving methodology affecting researchers and practitioners in areas ranging from science, technology, business and commerce. Integrated intelligent systems are gaining better acceptance in engineering design. The driving force behind this is that integrated intelligence and distributed 3C (collaboration, cooperation, and coordination) allows the capture of human knowledge and the application of it so as to achieve high-quality designs/products. Further motivation arises from steady advances in individual and hybrid intelligent-systems techniques, and the widespread availability of computing resources and communications capability through intranets and the web.

There is a need for an edited collection of articles to reflect emerging integrated intelligent technologies and their applications in engineering design. The great breadth and expanding significance of AI and integrated intelligent systems (IIS) fields on the international scene requires a major reference work for an adequately substantive treatment of the subject. It is intended that this work will fulfill this need.

The Objective of the Book

This book aims to describe recent findings and emerging techniques that use intelligent systems (particularly integrated and hybrid paradigms) in engineering design, and examples of applications. The goal is to take a snapshot of progress relating to research into systems for supporting design and to disseminate the way in which recent developments in integrated, knowledge-intensive, and computational AI techniques can improve and enhance such support. The selected articles provide an integrated, holistic perspective on this complex set of challenges and provide rigorous research results. The focus of this book is on the integrated intelligent methodologies, frameworks and systems for supporting engineering design activities. The subject pushes the boundaries of the traditional topic of engineering design into new areas.

The Target Audience of the Book

We intend this book to be of interest to researchers, graduate students and practicing engineers involved in engineering design and applications using integrated intelligent techniques. In addition, managers and others can use it to obtain an overview of the subject, and gain a view about the applicability of this technology to their business. As AI and intelligent systems technologies are fast evolving, we certainly hope that this book can serve as a useful insight to the readers on the state-of-the-art applications and developments of such techniques at the time of compilation.

The Organization of the Book

The chapters provide an integrated, holistic perspective on the complex set of challenges, combined with practical experiences of leading experts in industry. Some of the chapters provide rigorous research results, while others are in-depth reports from the field. All chapters have been rigorously reviewed and carefully edited. There is a logical flow through this book, starting with intelligence foundations, emerging intelligent techniques, frameworks, systems and tools then continuing integrated and hybrid intelligent systems followed by their applications for engineering design. The treatment of the subject in the book can be described as:

1) Examines emerging technologies and recent research results on AI and integrated intelligent systems (IIS) in engineering design, including integrated intelligent systems, soft computing, distributed artificial intelligence (DAI), computer-integrated information systems (CIIS), etc.

2) Introduces new knowledge-intensive problem-solving strategies and their implementations based on AI and integrated intelligent systems techniques.

3) Presents theoretical fundamental principles and implementation technologies as well as engineering design applications and case studies, including, for example, electro-mechanical assemblies and systems, process control system, embedded and mechatronic systems design.

This book consists of 20 chapters organized into three thematic sections. An overview of each section and a brief description of the component chapters are presented here.

Part I: Intelligence Foundations for Engineering Design. This section, consisting of Chapters 1 to 5, provides the theoretical foundations of specific AI and IIS-based technologies for engineering design, including the principles of directed mutations for evolutionary algorithms, fuzzy logic and many valued logic, swarm intelligence, constraint satisfaction problem, fuzzy set and logic, fuzzy linear programming, Bayesian model, decision tree, uncertainty, etc.

Chapter 1, by Stefan Berlik and Bernd Reusch, introduces directed mutation as well as different operators in one single place. Their characteristics such as a multivariate skew distribution as a mutation operator in a covariance matrix adaptation algorithm are presented. An application scenario and experimental results solving a real world optimization task in this scenario are presented to show how evolutionary algorithms and directed mutation can be applied in engineering design.

Chapter 2, by Kalle Saastamoinen, studies the properties and usability of basic many valued-structures known as t-norms, t-conorms, implications and equivalences in comparison tasks. It shows how these measures can be aggregated with generalized mean and what kind of measures for comparison can be achieved from this procedure.

Chapter 3, by Arun Khosla, Shakti Kumar, K. Aggarwal, and Jagatpreet Singh, reports a swarm intelligence (SI) technique, Particle Swarm Optimization (PSO), which is a robust stochastic evolutionary computation engine. This is emerging as an innovative and powerful computational metaphor for solving complex problems in design, optimization, control, management, business and finance. The focus of this chapter is to present the use of the PSO algorithm for building optimal fuzzy models from the available data in design of the rapid Nickel-Cadmium (Ni-Cd) battery charger.

Chapter 4, by Arijit Bhattacharya and Pandian Vasant, outlines an intelligent fuzzy linear programming (FLP) method that uses a flexible logistic membership function (MF) to determine fuzziness patterns at disparate level of satisfaction for theory of constraints (TOC) based product-mix design problems. The fuzzy-sensitivity of the decision has been focused for a bottle-neck-free, optimal product-mix solution of TOC problem.

Chapter 5, by Vitaly Schetinin, Jonathan Fieldsend, Derek Partridge, Wojtek, Krzanowski, Richard Everson, Trevor Bailey, and Adolfo Hernandez, proposes a new approach to decision trees (DTs) for the Bayesian Markov Chain Monte Carlo technique to estimate uncertainty of decisions in safety-critical engineering applications. It also proposes a new procedure of selecting a single DT and describes an application scenario.

Part II: Techniques, Frameworks, Tools and Standards. This section, containing Chapters 6 to 11, explores techniques, models and frameworks, both current and emerging, and potential architectures for intelligent integrated engineering design.

Chapter 6, by Xiang Li, Junhong Zhou, and WenFeng Lu, presents a set of customer requirement discovery methodologies to achieve broad and complex market studies for new products. The proposed approach uses data mining and text mining technologies to discover customer multi-preference and corresponding customer motivation. A prototype system that allows for on-line customer feedback collection, digitization of the language feedbacks, numerical descriptions of customer preferences and customer motivation of a product is developed to demonstrate the feasibility of the proposed methodologies. It is shown that the proposed work could significantly shorten the survey and analysis time for customer preference and is thus expected to help companies to reduce design cycle time for new product design.

Chapter 7, by Paulo Gomes, Nuno Seco, Francisco Pereira, Paulo Paiva, Paulo Carreiro, José Ferreira and Carlos Bento, introduces an approach to reusing the knowledge gathered in the design phase of software development. An intelligent CASE tool using case-based reasoning (CBR) techniques and WordNet is developed to support software design and provide a framework for storage and reuse of design knowledge. This Chapter presents the approach to exploiting a knowledge base and several reasoning mechanisms that reuse the stored knowledge.

Chapter 8, by W.D. Li, S.K. Ong, A.Y.C., Nee, L. Ding, and C.A. Mcmahon, proposes and develops three intelligent optimization methods, i.e., Genetic Algorithm (GA), Simulated Annealing (SA) and Tabu Search (TS). These are applied to the solution of intractable decision-making issues in process planning with complex machining constraints. These algorithms can determine the optimal or near-optimal allocation of machining resources and sequence of machining operations for a process plan simultaneously, and a fuzzy logic-based Analytical Hierarchical Process technique is applied to evaluate the satisfaction degree of the machining constraints for the process plan.

Chapter 9, by Andrew Feller, Teresa Wu, and Dan Shunk, reviews existing research and industry-based practices relating to collaborative product design (CPD) information systems. An information framework is proposed called the 'Virtual Environment for Product Development' (VE4PD) that is based on the integration of Web services and agent technologies to manage the CPD process. The VE4PD architecture is designed to support CPD functions such as design synchronization, timely notification, distributed control, role based security, support for distributed intelligent agents, and varying design rule standards. An implementation system including intelligent agents for design negotiation is also described that validates the application approach.

Chapter 10, by Zhu Fan, Mogens Andreasen, Jiachuan Wang, Erik Goodman, and Lars Hein, proposes an integrated evolutionary engineering design framework that integrates the chromosome model in the domain theory, the evolutionary design, and human interaction. The evolvable chromosome model can help the designer to improve creativity in the design process, suggesting to them unconventional design concepts, and preventing them from looking for solutions only in a reduced solution space. The systematic analytical process to obtain a chromosome model followed by evolutionary design algorithms also helps the designer to have a complete view of design requirements and intentions. Human interaction is integrated to the framework due to the complex and dynamic nature of engineering design. It also helps the designer to accumulate design knowledge and form a design knowledge base. An example of the design of a vibration absorber for a typewriter demonstrates the feasibility of the technique.

Chapter 11, by Xuan F. Zha, proposes an integrated intelligent approach and a multi-agent framework for the evaluation of the assemblability and assembly sequence of electro-mechanical assemblies (EMAs). The proposed approach integrates the STEP (STandard for the Exchange of Product model data, officially ISO 10303) based assembly model and XML schema with a fuzzy analytic hierarchy process. Through integration with the STEP-based product modeling agent system, a CAD agent system and assembly planning agent system, the developed assembly evaluation agent system can effectively incorporate, exchange, and share concurrent engineering knowledge into the preliminary design process so as to provide users with suggestions for improving a design and also helping obtain better design ideas.

Part III: Applications. Chapters 12 to 20 in this section address the important issue of the ways that integrated intelligent systems are applied in engineering design. Case studies examine a wide variety of application areas including benchmarking and comparative analysis. The basic question is how accumulated data and expertise from engineering and business operations can be abstracted into useful knowledge, and how such knowledge can be applied to support engineering design. In this part of the book, chapters report case studies of innovative actual IIS-ED applications deploying specific AI-based technologies, such as logic rule-based systems, neural networks, fuzzy logic, cased-based reasoning, genetic algorithms, data mining algorithms, intelligent agents, and user intelligent interfaces, among others, and the integrations of these paradigms.

Chapter 12, by Sarawut Sujitjorn, Thanatchai Kulworawanichpong, Deacha Puang-downrecong and Kongpan Areerak,presents detailed step-by-step description of an intelligent search algorithm known as 'Adaptive Tabu Search' (ATS). The proof of its convergence and its performance evaluation are illustrated. The chapter demonstrates the effectiveness and usefulness of the ATS through various engineering applications and designs in the following fields: power system, identification, and control.

Chapter 13, by Shi-Shang Jang, David Shun-Hill Wong and Junghui Chen, addresses a technique known as 'experimental design' describes The design of new processes in modern competitive markets is mainly empirical because the short life-cycle does not allow the development of first-principle models. A systematic methodology know as 'experimental design', based on statistical data analysis and decision making, is used to optimise the number of experiments and direct process development. However, such methods are unsatisfactory when the number of design variables becomes very large and there are non-linearities in the input-output relationship. The new approach described in this chapter uses artificial neural networks as a meta-model, and a combination of random-search, fuzzy classification, and information theory as the design tool. An information free energy index is developed which balances the needs for resolving the uncertainty of the model and the relevance to finding the optimal design. The procedure involves iterative steps of meta-model construction, designing new experiments using meta-model and actual execution of designed experiments. The effectiveness of this approach is benchmarked using a simple optimization problem. Three industrial examples are presented to illustrate the applicability of the method to a variety of design problem.

Chapter 14, by Miki Fukunari and Charles J. Malmborg, proposes computationally efficient cycle time models for Autonomous Vehicle Storage and Retrieval System that use scalable computational procedures for large-scale design conceptualization. Simulation based validation studies suggest that the models produce high accuracy. The procedure is demonstrated for over 4,000 scenarios corresponding to enumeration of the design spaces for a range of sample problems.

Chapter 15, by Hirotaka Nakayama, Koichi Inoue and Yukihiro Yoshimori, discusses approximate optimization methods developed using computational intelligence, in which optimization is performed in parallel with the prediction of the form of the objective function. In this chapter, radial basis function networks (RBFN) are employed in predicting the form of objective function, and genetic algorithms (GA) used in searching for the optimal value of the predicted objective function. The effectiveness of the suggested method is shown through some numerical examples along with an application to seismic design in reinforcement of cablestayed bridges.

Chapter 16, by Glenn Semmel, Steven R. Davis, Kurt W. Leucht, Daniel A. Rowe, Kevin E. Smith, Ladislau Bölöni, discusses a rule-based telemetry agent used for Space Shuttle ground processing. It presents the problem domain along with design and development considerations such as information modeling, knowledge capture, and the deployment of the product. It also presents ongoing work with other condition monitoring agents.

Chapter 17, by Mitun Bhattacharyya, Ashok Kumar, Magdy Bayoumi, proposes two techniques in two different sub areas of Wireless Sensor Networks (WSN) to reduce energy using learning methods. In the first technique, a watch-dog/blackboard mechanism is introduced to reduce query transmissions, and a learning approach is used to determine the query pattern from the cluster head. Once the pattern is learnt, data are automatically sent back even in the absence of queries from the cluster head. In the second technique a learning agent method of profiling the residual energies of sensors within a cluster is proposed.

Chapter 18, Juan Vidal, Manuel Lama, and Alberto Bugarin, describes a knowledge-based system approach that combines problem-solving methods, workflow and machine learning technologies for dealing with the furniture estimate task. The system integrates product design in a workflow-oriented solution, and is built over a workflow management system that delegates the execution of activities to a problem-solving layer. An accurate estimation of the manufacturing cost of a custom furniture client order allows competitive prices, better profits adjustment, and increments the client portfolio too.

Chapter 19, by Martin Böhner, Hans Holm Frühauf and Gabriella Kókai,discusses the suitability of ant colony optimization (ACO) to an employment with blind adaptation of the directional characteristic of antenna array systems. In order to fulfill the hard real time constraints for beam forming in ranges of few milliseconds a very efficient hardware implementation for a highly parallel distributed logic is proposed in this chapter. The application requirements are given because of the high mobility of wireless subscribers in modern telecommunication networks. Such a dynamic alignment of the directional characteristic of a base-station antenna can be achieved with the help of a hardware-based Ant Colony Optimization methodology, by controlling the steering antenna array system parameters as digital phase shifts and amplitude adjustment. By means of extensive simulations it was confirmed that the suggested ACO fulfills the requirements regarding the highly dynamic changes of the environment. Based on these results a concept is presented to integrate the optimizing procedure as high-parallel digital circuit structure in a customized integrated circuit of a reconfigurable gate array.

Chapter 20, by Ankur Agarwal, Ravi Shankar, A. S. Pandya, presents the application of genetic algorithms to system level design flow to provide best effort solutions for two specific tasks, viz., performance tradeoff and task partitioning. Multiprocessor system on chip (MpSoC) platform has set a new innovative trend for the system-on-chip (SoC) design. Demanding Quality of Service (QOS) and performance metrics are leading to the adoption of a new design methodology for MpSoC. These will have to be built around highly scalable and reusable architectures that yield high speed at low cost and high energy efficiency for a variety of demanding applications. Designing such a system, in the presence of such aggressive QOS and Performance requirements, is an NP-complete problem.

Summary

There are over 48 coauthors of this notable work and they come from 19 countries. The chapters are clearly written, self-contained, readable and comprehensive with helpful guides including introduction, summary, extensive figures and examples and future trends in the domain with comprehensive reference lists. The discussions in these parts and chapters provide a wealth of practical ideas intended to foster innovation in thought and consequently, in the further development of technology. Together, they comprise a significant and uniquely comprehensive reference source for research workers, practitioners, computer scientists, academics, students, and others on the international scene for years to come.

Acknowledgment

The original intention of this edited book came from the discussion on the proposition to publish a Special Issue on Integrated and Hybrid Intelligent Systems in Product Design and Development in the International Journal of Knowledge-based and Intelligent Engineering Systems (KES) (Bob is Chief Editor of KES). The special issue was out in June 2005. We thought that based on the Special Issue in KES the selected quality papers could be extended into chapters, supplementing with additional chapters and forming the whole into a book that would fit well into the knowledge-based intelligent engineering systems collection of IOS Press. We then started working together in this direction. We were quite excited about this movement and immediately contacted the contributors and spread call for papers. The response from all the contributors was very positive and the proposal for a book was submitted to IOS for evaluation. The good news that the IOS Editorial Committee had approved the publishing of the book was conveyed to us in August 2005. We were overjoyed that the call-for-papers of the Special Issue and chapters had attracted favorable responses from many top researchers in this field. As the original intention was a peer reviewed Special Issue, and all the papers were either in the process of being reviewed or had already gone through the reviewing process, we informed the contributors that the quality of each paper, now each chapter, had followed the same standard of a rigorously peer-reviewed international journal.

We are most grateful to the kind cooperation of all the contributors who had promptly responded to all the questions and had followed our requests for additional information. We would also like to thank IOS for giving us this opportunity of publishing this book.

Xuan F. (William) Zha, Gaithersburg, Maryland

Robert J. (Bob) Howlett, Brighton, UK

Directed mutation abandons the so-called random mutation hypothesis postulating mutations to occur at random, regardless of fitness consequences to the resulting offspring. By introducing skewness into the mutation operators, bigger portions of offspring can be created in the area of higher fitness with respect to the elder and thus promising directions of the evolution path can be favored. The aim of this work is to present the foundations of directed mutation as well as different operators in one single place. Their characteristics will be presented and their advantages and disadvantages are discussed. Furthermore, an application scenario will be presented that shows how evolutionary algorithm and directed mutation can be applied in engineering design. In addition, some experimental results solving a real world optimization task in this scenario are provided. Finally some first, preliminary results of a multivariate skew distribution as mutation operator in a covariance matrix adaptation algorithm will be presented.

In this chapter we will study properties and usability of basic many valued structures called t-norms, t-conorms, implications and equivalences in comparison tasks. We will show how these measures can be aggregated with generalized mean and what kind of measures for comparison can be achieved from this procedure. New classes for comparison measures are suggested, which are combination measure based on the use of t-norms and t-conorms and pseudo equivalence measures based on S-type implications.

In experimental part of this chapter we will show how some of the comparison measures presented here work in comparison task. For comparison task we use classification. We show by comparison to results that can be achieved through some known public domain classifier results that our classification results are highly competitive.

Particle Swarm Optimization (PSO), which is a robust stochastic evolutionary computation engine, belongs to the broad category of swarm intelligence (SI) techniques. SI paradigm has been inspired by the social behavior of ants, bees, wasps, birds, fishes and other biological creatures and is emerging as an innovative and powerful computational metaphor for solving complex problems in design, optimization, control, management, business and finance. SI may be defined as any attempt to design distributed problem-solving algorithms that emerges from the social interaction. The objective of this chapter is to present the use of PSO algorithm for building optimal fuzzy models from the available data. The fuzzy model identification procedure using PSO as an optimization engine has been implemented as a Matlab toolbox and is also presented in this chapter. For the purpose of illustration and validation of the approach, the data from the rapid Nickel-Cadmium (Ni-Cd) battery charger developed by the authors has been used.

This chapter outlines an intelligent fuzzy linear programming (FLP) having a flexible logistic membership function (MF) in finding out fuzziness patterns at disparate level of satisfaction for theory of constraints-based (TOC) product-mix design problems. One objective of the present work is to find out degree of fuzziness of product-mix decisions having disparate level of satisfaction of decision-maker (DM). Another objective is to provide a robust, quantified monitor of the level of satisfaction of DMs and to calibrate these levels of satisfaction against DMs' expectations. Fuzzy-sensitivity of the decision has been focused for a bottle-neck-free, optimal product-mix solution of TOC problem.

Uncertainty of decisions in safety-critical engineering applications can be estimated on the basis of the Bayesian Markov Chain Monte Carlo (MCMC) technique of averaging over decision models. The use of decision tree (DT) models assists experts to interpret causal relations and find factors of the uncertainty. Bayesian averaging also allows experts to estimate the uncertainty accurately when a priori information on the favored structure of DTs is available. Then an expert can select a single DT model, typically the Maximum a Posteriori model, for interpretation purposes. Unfortunately, a priori information on favored structure of DTs is not always available. For this reason, we suggest a new prior on DTs for the Bayesian MCMC technique. We also suggest a new procedure of selecting a single DT and describe an application scenario. In our experiments on real data our technique outperforms the existing Bayesian techniques in predictive accuracy of the selected single DTs.

Effective collection and analysis of customer demand is a critical success factor for new product design and development. This chapter presents a set of customer requirement discovery methodologies to achieve broad and complex market studies for new products. The proposed approach uses data mining and text mining technologies to discover customer multi-preference and corresponding customer motivation. Using the proposed rule mining methodology, discovery rules can be flexibly defined, the complete customer multi-preference patterns are discovered and their statistic analysis of multi-preference can be conducted for new product design. With the proposed text mining methodology, the customer motivations are discovered and the percentage of surveyed customers with certain preference and the reason for this preference are presented. Combining the methodologies in text mining with rule mining, the customer motivations can be quantitatively described with statistic analysis results. A prototype system that allows on-line customer feedback collection, digitization of the language feedbacks, numerical descriptions of customer preferences and customer motivation of a product is developed to demonstrate the feasibility of the proposed methodologies. It is shown that the proposed work could significantly shorten the survey and analysis time for customer preference and is thus expected to help companies to reduce cycle time for new product design.

Reusing the knowledge gathered in the design phase of software development is an important issue for any software company. It enables software developers to work faster and make fewer mistakes, which decreases the development time due to the increased efficiency of the development team. In order to accomplish design knowledge reuse, we have developed an intelligent CASE tool that supports software design. Our system uses Case-Based Reasoning and WordNet, providing a framework for storage and reuse of design knowledge. This chapter presents our approach, which exploits a knowledge base and several reasoning mechanisms that reuse the stored knowledge.

Manufacturing cost is crucial for the economic success of a product, and early and accurate estimation of manufacturing cost can support a designer to evaluate a designed model dynamically and efficiently for making cost-effective decisions. Manufacturing cost estimation is closely related to process planning problems, in which machining operations, machining resources, operation sequences, etc., are selected, determined and optimized. To solve the intractable decision-making issues in process planning with complex machining constraints, three intelligent optimization methods, i.e., Genetic Algorithm (GA), Simulated Annealing (SA) and Tabu Search (TS), have been developed to determine the optimal or near-optimal allocation of machining resources and sequence of machining operations for a process plan simultaneously, and a fuzzy logic-based Analytical Hierarchical Process technique has been applied to evaluate the satisfaction degree of the machining constraints for the process plan. Case studies, which are used to compare the three developed methods, are discussed to highlight their characteristics in the aspects of solution quality, computation efficiency and optimization result robustness.

Improving and streamlining the communication and decision processes used in Collaborative Product Development (CPD) requires a robust Distributed Information System that can enable intelligent notification, coordination and negotiation processes. In this chapter we review existing research and industry practice related to CPD information systems and propose an information framework called the Virtual Environment for Product Development (VE4PD) that is based on the integration of Web services and agent technologies to manage the CPD process. The VE4PD architecture is designed to support CPD functions such as design synchronization, timely notification, distributed control, role based security, support for distributed intelligent agents, and varying design rule standards. An implementation system including intelligent agents for design negotiation is also described that validates the application approach.

The paper proposes an integrated evolutionary engineering design framework that integrates the chromosome model in the domain theory, the evolutionary design, and human interaction. The evolvable chromosome model helps the designer to improve creativity in the design process, suggesting them with unconventional design concepts, and preventing them from looking for solutions only in a reduced solution space. The systematic analytical process to obtain a chromosome model before running evolutionary design algorithms also helps the designer to have a complete view of design requirements and intentions. Human interaction is integrated to the framework due to the complex and dynamic nature of engineering design. It also helps the designer to accumulate design knowledge and form a design knowledge base. An example of vibration absorber design for a typewriter demonstrates its feasibility.

Assemblability analysis and evaluation plays a key role in assembly design, operation analysis and planning. In this paper, we propose an integrated intelligent approach and framework for evaluation of assemblability and assembly sequence for electro-mechanical assemblies (EMAs). The approach integrates the STEP (STandard for the Exchange of Product model data, officially ISO 10303) based assembly model and XML schema with the fuzzy analytic hierarchy process for assembly evaluation. The evaluation structure covers not only the geometric and physical characteristics of the assembly parts but also the assembly operation data necessary to assemble the parts. The realization of the integration system is implemented through a multi-agent framework. Through integration with the STEP-based product modeling agent system, CAD agent system and assembly planning agent system, the developed assembly evaluation agent system can effectively incorporate, exchange, and share concurrent engineering knowledge into the preliminary design process so as to provide users with suggestions for improving a design and also helping obtain better design ideas. The proposed approach has the flexibility to be used in various assembly methods and different environments. The applications show that the proposed approach and system are feasible.

This chapter presents detailed step-by-step description of an intelligent search algorithm namely Adaptive Tabu Search (ATS). The proof of its convergence, and its performance evaluation are illustrated. The chapter demonstrates the effectiveness and usefulness of the ATS through various engineering applications and designs in the following fields: power system, identification, and control.

Ability to rapidly design products and their manufacturing process is a key to being competitive in a dynamic market environment. Traditional methods of design of experiment development are unsatisfactory when applied to design problems with large number of input variables and nonlinear input-output relation. A meta-model driven experimental design scheme is developed. The approach uses artificial neural network as the meta-model, and a combination of random-search, fuzzy classification, and information theory as the design tool. An information free energy index is developed which balances the needs for resolving the uncertainty of the model and the relevance to finding the optimal design. The procedure involves iterative steps of meta-model construction, designing new experiments using meta-model and actual execution of designed experiments. The effectiveness of this approach is benchmarked using a simple optimization problem. Three industrial examples are presented to illustrate its applicability to a variety of design problem.

Unit load storage systems are pervasive throughout global supply chains. Significant reductions in their automation costs could have significant economic impact. A new alternative to traditional crane-based automation uses flexible autonomous vehicles in storage and retrieval operations. This technology has not significantly penetrated commercial markets for warehouse automation due to a lack of design tools for evaluating its performance. This precludes direct comparisons of autonomous vehicle technology with crane-based technology in the pre-engineering or “design conceptualization” stage of system development where key technology selection decisions are made. To address this problem, this chapter proposes computationally efficient cycle time models for Autonomous Vehicle Storage and Retrieval System that use scalable computational procedures for large-scale design conceptualization. Simulation based validation studies suggest that the models produce high accuracy. The procedure is demonstrated for over 4,000 scenarios corresponding to enumeration of the design spaces for a range of sample problems.

In many practical engineering design problems, the form of objective function is not given explicitly in terms of design variables. Given the value of design variables, under this circumstance, the value of objective function is obtained by some analysis such as structural analysis, fluidmechanic analysis, thermodynamic analysis, and so on. Usually, these analyses are considerably time consuming to obtain a value of objective function.

In order to make the number of analyses as few as possible, approximate optimization methods using computational intelligence have been developed. In those methods, optimization is performed in parallel with predicting the form of objective function. In this paper, radial basis function networks (RBFN) are employed in predicting the form of objective function, and genetic algorithms (GA) in searching the optimal value of the predicted objective function. One of the most important tasks in this approach is to allocate sample data moderately in order to make the number of experiments as small as possible. The effectiveness of the suggested method will be shown through some numerical examples along with an application to seismic design in reinforcement of cable-stayed bridges.

The Engineering Development Directorate at NASA Kennedy Space Center has designed, developed, and deployed a rule-based agent to monitor the Space Shuttle's ground processing telemetry stream. The NASA Engineering Shuttle Telemetry Agent increases situational awareness for system and hardware engineers during ground processing of the Shuttle's subsystems. The agent provides autonomous monitoring of the telemetry stream and automatically alerts system engineers when user defined conditions are satisfied. Efficiency and safety are improved through increased automation.

Sandia National Labs' Java Expert System Shell is employed as the agent's rule engine. The shell's predicate logic lends itself well to capturing the heuristics and specifying the engineering rules within this domain. The declarative paradigm of the rule-based agent yields a highly modular and scalable design spanning multiple subsystems of the Shuttle. Several hundred monitoring rules have been written thus far with corresponding notifications sent to Shuttle engineers. This chapter discusses the rule-based telemetry agent used for Space Shuttle ground processing. We present the problem domain along with design and development considerations such as information modeling, knowledge capture, and the deployment of the product. We also present ongoing work with other condition monitoring agents.

In this chapter we propose two techniques in two different sub areas of Wireless Sensor Network (WSN) to reduce energy using learning methods. In the first technique we introduce a watchdog/blackboard mechanism to reduce query transmissions, in which an approach is used to learn the query pattern from cluster head. Once the pattern is learnt, data are automatically sent back even without any queries from the cluster head. In the second technique we propose a learning agent method of profiling the residual energies of sensors within a cluster. Here as the agent moves across the different sensor nodes, it profiles the sensors' residual energies. This profile information is provided to sensors to help them make intelligent routing decisions.

This chapter describes a knowledge-based system approach that combines problem-solving methods, workflow and machine learning technologies for dealing with the furniture estimate task. The system integrates product design in a workflow-oriented solution, and is built over a workflow management system that delegates activities execution to a problem-solving layer. An accurate estimation of the manufacturing cost of a custom furniture client order allows competitive prices, better profits adjustment, and increments the client portfolio too. Nevertheless, task scope is even broader. On one hand, it fixes future material and storage capacity requirements. On the other hand, it defines the manufacturing plan and logistic requirements to fulfil the client order in time. However, these objectives cannot be achieved without an adequate product design, which relates client order requirements with a manufacturing and assembly-oriented design.

The following chapter describes and discusses the suitability of ant colony optimization (ACO) to an employment with blind adaptation of the directional characteristic of antenna array systems. Due to the special advantage of the ACOs their robustness and high adaptation rate - this very young heuristically optimization strategy allows its application in the field of high radio frequency systems. In order to fulfil the hard real time constraints for beam forming in ranges of few milliseconds a very efficient hardware implementation for a highly parallel distributed logic is proposed in this chapter. The application requirements are given because of the high mobility of wireless subscribers in modern telecommunication networks. Such a dynamic alignment of the directional characteristic of a base-station antenna can be achieved with the help of a hardware based Ant Colony Optimization, by controlling the steerable antenna array system parameters as digital phase shifts and amplitude adjustment. By means of extensive simulations it was confirmed that the suggested ACO fulfils the requirements regarding the highly dynamic changes of the environment. Based on these results a concept is presented to integrate the optimizing procedure as high-parallel digital circuit structure in a customized integrated circuit of a reconfigurable gate array.

Multiprocessor system on chip (MpSoC) platform has set a new innovative trend for the system-on-chip (SoC) design. Demanding Quality of Service (QOS) and performance metrics are leading to the adoption of a new design methodology for MpSoC. These will have to be built around highly scalable and reusable architectures that yield high speed at low cost and high energy efficiency for a variety of demanding applications. Designing such a system, in the presence of such aggressive QOS and Performance requirements, is an NP-complete problem. In this paper, we present the application of genetic algorithms to system level design flow to provide best effort solutions for two specific tasks, viz.., performance tradeoff and task partitioning.