To fully exploit VLSI technology, future CAD tools need to perform a high proportion of the intelligent decision-making necessary for design. It is imperative that these tasks be massively automated so that the intellectual load on the human designer is reduced, allowing a more efficient use of his time and creative abilities. This book investigates the applicability of knowledge-based techniques to the problems of VLSI design. Topics covered include: an overview of knowledge-based techniques; a description of work illustrating the application of these techniques to various VLSI design tasks; and a survey of workstations and software available for constructing knowledge-based systems. Although much of the material in this book has been illustrated using examples from microelectronic CAD, the issues and techniques have been presented in such a way as to make them useful to managers, designers and researchers who wish to learn of the advantages of knowledge-based techniques in practical applications.

Samples the present state-of-the-art in CAD for VLSI, covering both newly developed algorithms and applications of techniques from the artificial intelligence community. The material is based on a tutorial course run in conjunction with the 1991 European Conference on Circuit Theory and Design, and should interest engineers involved in the design and testing of integrated circuits and systems. Annotation copyrighted by Book News, Inc., Portland, OR

This volume is about automation - automation in design, automation in manufacturing, and automation in production. Automation is essen tial for increased productivity of quality products at reduced costs. That even partial or piecemeal automation of a production facility can deliver dramatic improvements in productivity has been amply demon strated in many a real-life situation. Hence, currently, great ef forts are being devoted to research and development of general as well special methodologies of and tools for automation. This volume re ports on some of these methodologies and tools. In general terms, methodologies for automation can be divided into two groups. There are situations where a process, whether open-loop or closed-loop, is fairly clearly understood. In such a situation, it is possible to create a mathematical model and to prescribe a mathe matical procedure to optimize the output. If such mathematical models and procedures are computationally tractable, we call the correspond ing automation - algorithmic or parametric programming. There is, however, a second set of situations which include process es that are not well understood and the available mathematical models are only approximate and discrete. While there are others for which mathematical procedures are so complex and disjoint that they are computationally intractable. These are the situations for which heuristics are quite suitable for automation. We choose to call such automation, knowledge-based automation or heuristic programming.

The importance of CAD to electronics technology Computer-aided design (CAD) is one way of coping with the problem of how to design and build very complex systems. This problem is particularly acute in electronics technology. Designs are now (1984) said to be design-limited, rather than technology-limited. It can take months to generate a design for a chip, so that it might be obsolete before it can be manufactured. Manual design of large-scale integration (LSI) chips (circa 10,000 gates) is almost impossible. However, using current technology it is possible to produce chips having 250,000 gates. It is understandable, therefore, that there is great interest in improving existing CAD systems. Designers of CAD systems are concerned with formalizing and automating as much of the design task as possible. Automating design of any kind has long been acknowledged as a project fraught with intractable problems. A human designer has to have an understanding of the nature of the materials used in manufacture, a knowledge of common problems and well-tried solutions, and above all, creativity in producing new designs. Understanding, knowledge and creativity are three properties even the most artificially intelligent of computer programs have been entirely lacking in until very recently. Some people would deny computers these qualities entirely, on philosophical grounds (eg Dreyfus 1979; Searle 1981). There are few theories in cognitive psychology which can help.

Innovation in Product Design gives an overview of the research fields and achievements in the development of methods and tools for product design and innovation. It presents contributions from experts in many different fields covering a variety of research topics related to product development and innovation. Product lifecycle management, knowledge management, product customization, topological optimization, product virtualization, systematic innovation, virtual humans, design and engineering, and rapid prototyping are the key research areas described in the book. It also details successful case studies developed with industrial companies. Innovation in Product Design is written for academic researchers, graduate students and professionals in product development disciplines who are interested in understanding how novel methodologies and technologies can make the product development process more efficient.

As the cost of developing new semiconductor technology at ever higher bit/gate densities continues to grow, the value of using accurate TCAD simu lation tools for design and development becomes more and more of a necessity to compete in today's business. The ability to tradeoff wafer starts in an advanced piloting facility for simulation analysis and optimization utilizing a "virtual fab" S/W tool set is a clear economical asset for any semiconductor development company. Consequently, development of more sophisticated, accurate, physics-based, and easy-to-use device and process modeling tools will receive continuing attention over the coming years. The cost of maintaining and paying for one's own internal modeling tool development effort, however, has caused many semiconductor development companies to consider replacing some or all of their internal tool development effort with the purchase of vendor modeling tools. While some (noteably larger) companies have insisted on maintaining their own internal modeling tool development organization, others have elected to depend totally on the tools offered by the TCAD vendors and have consequently reduced their mod eling staffs to a bare minimal support function. Others are seeking to combine the best of their internally developed tool suite with "robust", "proven" tools provided by the vendors, hoping to achieve a certain synergy as well as savings through this approach. In the following sections we describe IBM's internally developed suite of TCAD modeling tools and show several applications of the use of these tools.

The complete shop floor automation - a "lights out factory", where workers initially set up all machines, turn off the lights, lock the door and the machine churns up the parts - remains an unfulfilled dream. Yet when we look at the enormity of the process of automation and integration even for the most simply conceived part factory, we can recognize that automation has been applied and is being applied, more so when it made sense from a cost/benefit standpoint. It is our nature to be dissatisfied with near term progress, but when we realize how short a time the tools to do that automation have been available, the progress is clearly noteworthy - considering the multitudes of factors and the environment we have to deal with. Most of the automa tion problems we confront in today's environment are multidisciplinary in nature. They require not just the knowledge and experience in various distinct fields but good cooperation from different disci plined organizations to adequately comprehend and solve such problems. In Volume III we have many examples that reflect the current state of the art techniques of robotics and plant automation. The papers for Volume III have been arranged in a logical order of automation planning, automated assembly, robot programming and simula tion, control, motion coordination, communication and networking to factories of the future.

Knowledge-Intensive CAD clarifies and elaborates the concepts of knowledge-intensive design and CAD. In today's advanced manufacturing environment, CAD systems should not only assist designers and engineers during product design, but also in design information for use in later stages of the process such as production, distribution and operation. This book focuses on the sharing of knowledge across life-cycle stages and organizational boundaries.