TITLE : KNOWLEDGE-BASED SYSTEMS FOR MANAGEMENT DECISIONS.*

This book is about knowledge-based systems, a branch of artificial intelligence (AI). It gives an introduction to knowledge-based systems, and how to develop these systems in a wide range of business situations.

The book is intended for anyone-business manager or student of business management-interested in the subject. It will help readers learn how to develop knowledge-based systems for management decision making. It will also help them improve their general management decision-making skills, as well as specific job-related skills, in such areas as marketing, finance, operations, and accounting. No background in computers is needed for developing the systems described in this book.

The book should also be of interest to corporate managers looking for well-defined, simple ways to

. Encourage the development of knowledge-based systems throughout their organizations

. Get those systems used to improve operations

. Generally improve middle management decision-making skills

Those thinking about becoming knowledge engineers will also find the book useful. It covers one area of knowledge-based systems development in particular that is given little attention elsewhere-analyzing and reconstructing decision situations. The book is not, however, for computer scientists. It is an introductory book for managers, and so is not highly technical.

The book can be used for

. Courses in knowledge-based systems development

. Advanced courses in any business discipline-from accounting and finance to marketing, operations management, and business policy

. General management decision-making courses

. Training seminars for managers who want to learn knowledge-based systems development

. Individual managers interested in studying the area on their own

The book is based on courses given at the author's university and on seminars and consulting work done for industry.

If artificial intelligence is to have wide application in business, it is necessary to take the development initiative out of the hands of the computer technicians, and put it into the hands of business managers.

Experience has shown that it is easier in the long run to teach a manager how to develop a knowledge-based system, than it is to expect a computer technician to acquire the skills and experience of business experts. Whether this works in practice in your own company, of course, will depend on the competence and training of the individuals involved, and on the organizational atmosphere.

In mid-1985, for example, DuPont undertook a major knowlege-based systems development project. The project's organization was based on the premise that middle and lower level manager's best know how decisions are made in their jobs. Training and helping these managers acquire knowledge-based systems development skills appeared, then, to be the fastest and most effective way to put AI technology to work to make money for DuPont.

Within two years DuPont reported that hundreds of useful Al applications had been developed (or were under development) using this bottom up, hands-on approach. Experience then shows this to be the quickest way to get practical and useful knowledge-based applications that work to help run a company better. It would seem, therefore, that lower and middle level management have a great role to play in the AI boom.

Successful knowledge-based systems development efforts, such as DuPont's, have taught us a lesson. The rapid growth of AI applications in industry will come only from

. First, putting knowledge-based systems development tools and expertise in the hands of middle and lower level management

. Then, training and helping these managers to seek out and develop opportunities for knowledge-based systems development in their own jobs

This book is part of the effort to put the power of AI into the hands of these nontechnical managers.

Using this approach is especially important today because it is at the lower and middle management levels that shortages arising from the "baby bust" will be occurring over the next twenty years. These shortages are already showing up in the retail and service areas today. A wide range of knowledge-based systems will be needed to help fill this management gap.

The unusual aspect of knowledge-based systems, which distinguishes the area from other computer areas, is that 80 percent of the work involves understanding and modeling a manager's decision processes used in doing a specific management job. This helps explain why these managers are the most effective medium for building knowledge-based systems. These managers also benefit in another way from participating in developing these systems because it helps them learn how to do their jobs better.

Several aspects of the relationship between knowledge-based systems and decision making and problem solving in various business areas are discussed in this book:

1. The nature of knowledge-based systems? and how to develop them to support management planning and decision making in various business areas

2. The use of knowledge-based systems and their development to improve management decision-making skills in such business areas as marketing, accounting, finance, operations, and business policy development

The examples of knowledge-based systems given in this book fall into two categories.

1. Small prototype systems. A number of small prototype systems are described in detail in Part 2. This was done

. To introduce the reader to knowledge-based systems development technology and concepts

. To introduce the reader to applying this technology to developing knowledge-based systems in specific business areas

. To provide a basis for the study of the concept, structure, and design of larger systems

Within the context of these objectives, such small prototype systems are appropriate and useful. An experienced manager in any of the specific business areas involved may, however, find some of these systems somewhat trivial.

2. Larger knowledge-based systems. These are designed to give the reader a feeling for how larger systems work. These are not trivial systems. They are described only in general terms for several reasons:

. A detailed study of them would be beyond the introductory scope of this book

. Documenting the larger systems in detail would take as many pages as there are in this book

. They are proprietary systems that cost hundreds of thousands of dollars to develop

The book is divided into three parts. The first part introduces readers to knowledge-based systems, and how to develop them. The second part describes a variety of knowledge-based prototype systems that were developed by managers guided and assisted, by the author, for their own use on the job.

The third part, the Appendixes, gives examples of several systems developed by others and described in periodicals. In addition, the Appendixes provide some technical information on expert systems shells and several studies of knowledge engineering.

Comments and enquiries are welcome. For information about some advanced systems in these and related management decision-making areas or about any other aspect of this book, the reader should contact me through the publisher.

I wish to thank Steve Thompson of Micro Data Base Systems for his assistance in helping me learn GURU and develop some of the sample systems discussed in this book. Extensive development help was provided by Kenneth Chou, who was involved in the project from its inception. I also extend my thanks to the many managers, students, and fellow workers who provided help, including May-Mei Wong, Sweelim Chia, and Nancy Ward, and to my graduate assistants, including Yuan-I Lin and John Merseburg.

Substantial contributions and considerable assistance in developing the knowledge-based systems described in Part 2 was also provided by T. Connelly and E. Hagerty (Chapter 8); P. Sinaly, A. Barbera, and A. Wojcik (Chapter 9); B. Daly (Chapter 10); C. Zavala and G. Rosenfeld (Chapter 11); S. Memis (Chapter 12); L. First and P. Borocco (Chapter 13); E. Portnoy (Chapter 14); L. Lutzak and D. Chatman (Chapter 15); C. Ochs and J. Nelson (Chapter 16); J. Morison (Chapter 17); W. Holsten and D. Popper (Chapter 18); E. Conlon (Chapter 19); K. Zick and H. Gindin (Chapter 20); J. Merseburg (Chapter 22).

I wish to thank the following companies for furnishing software and other materials for research: Analytica Corp.; Ashton-Tate; Alacritous Inc.; Borland International; Decision Support Software Inc.; Digitalk Inc.; Experiences in Software Inc.; EXSYS Inc.; General Optimization Inc.; Human Intellect Systems; Intelligence Ware; Level Five Research; Lightyear Inc.; Meridian Education Corporation; Micro Data Base Systems Inc.; Microsoft Corporation; Neuron Data; Paperback Software International; Personal Computer Engineers; Programs In Motion; Reality Technologies Inc.; SPSS International; Texas Instruments Inc.

Last, I wish especially to thank D. G. Dologite who participated in a substantial way in all phases of the systems development work (and its underlying basic research) described in this book.