TECHNIQUES OF INFORMATION TECHNOLOGY —AN OBSERVATION ON EXPERT SYSTEMS


S.M. Salmat Ullah Bhuiyan* & Mohammed Shah Alam Chowdhury

1.  INTRODUCTION

Expert Systems are sometimes referred to as being a form of artificial intelligence. This is misleading as they are not systems which are capable of cialis canadian pharmacy thinking for themselves, instead they use and sift through human knowledge in the specific order in which they have been programmed. Therefore, expert systems can hardly be called artificial intelligence within the true meaning of the term. Expert systems are programmed with knowledge supplied by a person or a group of people with expertise on a particular topic.

cheapest levitra

These people are referred to as ‘experts’ because they have some specific body of knowledge which is not shared by the majority of other persons. Once their knowledge has been incorporated into it the system becomes an expert system. Thus an expert system is a computer programme which contains the accumulated knowledge or wisdom of an individual or group of individuals on a particular topic. The programme may be interrogated by the user to find answers to specific questions or problems.

2. TYPES OF EXPERT SYSTEMS

The expert systems have been classified according to the benefit which the organisation hopes to gain by installing them. They could have been classified according to the nature of the expert systems employed, viz. diagnostic systems, systems for selection and planning systems. However, the typiology of expert systems is briefly outlined below :

2.1 Expert Systems for Planning Expert systems can be used to plan a design or configure a product and Digital Equipment Corporation (DEC) was one of the first companies to use it this way1. Xcon, Xsel and Escort are the typical expert systems used for planning. Xcon checks the order and designs the layout of each computer, and is much more reliable than the human experts. Xsel is used by the salespersons in order to solve the problems of order backlog, order returns and renegotiation. Escort contains details of all the engineering knowledge about the basic operations and also the idiosyncracies of particular piece of plant2.

Clarks shoes (UK) uses an expert system for planning the different sizes, styles and colours to be made in batch production. The benefit to Clarks. is flexibility of production. This degree of flexibility could not be achieved if humans did the planning3.

2.2 Integrated Expert Systems

There   is  great  potential  with integration for reaping more benefits. One area in which progress   has already been made is by integrating CAD (Computer aided design) and expert systems to provide a ‘thinking’ system.   Progress  towards  fully linking CIM (Computer integrated manufacturing) and expert system is beginning4. Expert systems are available which operate in real time and which can monitor several thousands of variables at the same time. These systems can be used for process control.

2.3 Off-the Shelf Expert Systems

All of the systems considered up to now have been developed by an organization for internal use to solve a particular problem. It is possible to buy off-the shelf expert systems to give specific advice. British Legal publishers Butterworth Company developed  an  expert system  on ‘latent damage law‘. It is considered to be an obscure area of law of which few have much knowledge. This system  comes  together with  a booklet and consists of about a thousand rules. In comparison with the price  of legal books  this  is relatively quite cheap5.

2.4 Personal Productivity System

This is the simplest type of expert system.   It   has   low   levels   of complexity in both its embodied knowledge are technology. Examples include a personal budgeting system running on a PC built with a DOS-based expert systems shell. The key thrust of there systems is to improve personal  decision  making  and

* Associate Professor, Department of Marketing, University of Chittagong ” ” Lecturer in Marketing, Department of Marketing, University of Chittagong. thereby increase productivity6.

2.5 Power Decision Systems

This is knowledge intensive system and is relatively uncomplex in its technology. These systems incorporate the knowledge of highly skilled decision makers including the professionals working on difficult problems. Power Decision System (PDS) operates on relatively simple stand-alone computer. The PDSs are used for engineering analysis, financial and portfolio analysis, and medical diagnosis7.

2.6    Integrated Production Systems

These Systems involve advanced technology. These types of systems tend to target organizational productivity by improving throughput, reducing headcount. and lowering costs. Such systems might communicate regularly with larger administrative systems, access large database or be ported to a wide range of different computer hardware environments.

2.7 Strategic Impact Systems

The systems contain high level of complexity in its technology. They have multiple informations sources and the information is uncertain. The decision making process is long and intricate, and requires testing of numerous hypotheses. Strategic impact systems  often need high  levels of “systems integration”. Given the type of complexity and the unavoidable costs and time associated with it, a company must be very sure of the expected benefits of the system8. These benefits must come at different levels: improved decision making, organizational productivity, and greater marketing effectiveness.

3. Utilities of Expert Systems

Expert systems improve company image through more efficient services. They ensure the quality and consistency of decision making. They are used to concentrate on more critical problems. However, an organization may require an expert system for any of the following reasons :

3.1 Improved quality of decisions

The expert system can be programmed with the best knowledge available which will be based on the best expert working at the top of his/her form9. Thus, better decisions can be made regardless of the level of skill of the individual in charge of making them. Moreover, better qualify decisions save money providing the customer with a better and more consistent image of the organization.

3.2 Time Compression

An expert system can be used to save a considerable amount of time in the analysis of a problem and making the diagnosis. The amount of time taken for diagnosis is reduced to about one-tenth of that taken by humans10. Therefore, it seems reasonable to assume that level of reduction is possible in most instances.

2.3 Cost Savings

Considerable  cost savings  are possible through expert systems. Often, systems pay for themselves within the first year and most have recovered the initial outlay by the end of their second year of operation. This  saving in cost is generally achieved by speeding up the decision making process or by improving the quality of the decision11.

3.4 Training

There is plenty of potential of use expert systems to train staff both on and off the job. Expert systems can also be used to train staff in particular skills. Programmes can be used to put trainees through their paces by setting a problem and inviting the trainee to suggest a solution. If the answer is wrong the trainee can be taken through the decision stages step by step and given a printout on the correct decision    process12. Working alongside an expert system can also improve the knowledge of relatively inexperienced employee.

4. Criticism of Expert Systems

Expert systems can present a human dilemma. Knight and Silk mention that experts disclose their hard-won knowledge to a machine and thereby make themselves less essential13. It can be a way for a human expert to enshrine his knowledge in permanent form, and thus achieve some form of immortality. UK companies are most sceptical about the computer’s chances of taking over from human experts. Price Waterhouse survey reveals that, in the UK, 28% of companies have no faith in expert systems14. Moreover, a number of survey questionnaires were returned by the executives with the comment ” What is an expert system?” A number of expert systems are used for executive information system. Executive Information System in mostly confined to top management. By using EIS, the top executive sees what is going on but the manager below him may not have access to the same information even though that manager has his own critical success factors or performance indicators to monitor. It makes sense to provide similar information service right down the chain. Otherwise EIS will become known as Exclusive Information System (Bird).

5.  Conclusion

The best method of organizing an expert system depends entirely on the type of application. It could be argued that any company could attempt to build any one or all of the basic types of systems. In developing expert systems, human experts must be supported by knowledge engineers or system analyst. It is of course prudent to try to minimize technical complexity of the expert systems. Finally it can be said that expert systems or knowledge based systems will become a greater feature of most people’s lives whether in the factory, office or home.

References :

1.      Feigenbaum, E,; Mclcorduck, P.; & Nii. N.P.;  (1988).  the Rise of the  Expert Company, Macmillan.

2.      Barton, D.L. &SvloklaJ. (1988) “Putting Expert Systems  to  work”.   Harvard Business Review, Mar. /Apr.

3.      Cashmore L.& Lyall R (1991). Business Information Systems and Startigies.Prentice Hall, London. P. 199.

4.      MeyerMH&CurleyK.F(1991). “Putting Experts Systems Technology to Work” Sloan Management Review. Winter, PP-21—31.

5.      HarmonP, MansR&Morrisey W(1988). Expert Systems Tools and Applications, John Wiley & Sons, New York, P. 402.

6.      Harmon P. etal ibid.

7.      HartA. (1986) Knowledge Acquisition for Expert Systems. Me Graw-Hill, New York, P. 129.

8.      Clancy.    W.     (1985),     “Heruristic classification”, Artificial Intelligence, No. 27, PP. 289—350.

9.      Earl. MJ (1991) Management Startegies forlnformationTechnology, Perntice Hall, London, P. 79.

10.    O’Leary, D (1987),” Validation of Expert Systems”, Decision Sciences, Summer. PP. 468—486.

11.    Bobrow D, Mittal S. & Stefik M (1986) ” Expert Systems : Perils and Promise”, Communications of ACM, September PP. 880—894.

12.    Partridge D (1987)  ” The Scope  and Limitations of the First Generation of Expert Systems”,  Future  Generation Computer Systems. Vol: 3, No. 1 PP. 1— 10.

13.    Knight P; & Silk L (1990). Information Intensive Britain -A Critical Analysis of Policy issues,  Policy Study Institute. London, P-23.

14.    Managing Information— International Survey, 1988/89, Price Waterhouse, UK, PP. 10—14.

15.    BirdJane, (1992). “Managinglnformation Technology-Micro Myopia”, Management todya, February, PP. 102—109.

Subscribe / Share

Article by nuzhat akhter

Authors bio is coming up shortly.
nuzhat akhter tagged this post with: , , , , , , Read 35 articles by