Expert Systems In Prolog

Ian Lynch

Sunnybank State High School


Many secondary school course descriptions and syllabi now
provide for, and indeed encourage, a variety of teaching,
learning and assessment methods. Student project work is one
such vehicle at the disposal of teachers. A challenge facing
teachers is to define and implement criteria by  which project
work can be set, administered and assessed in accordance with
the teaching and computing environment in which they operate. 

An ideal focus for the core material of a student project, as
applied to the computer based solutions of problems, relates to
the four steps of Professor George Polya's problem solving
method. This approach is also known more generally as the
'Software Development Cycle'. Student project work, reflective
of this cycle, is a proven methodology which has been
associated with algorithmic problem solving strategies and
third generation computer languages in secondary schools for
the past two decades. The specific skills that are exercised by
this process are classified into the following groups.
1.   define the problem; 
2.   devise a plan to solve the problem;
3.   implement the plan by building a software prototype;
4.   execute test cases to debug the prototype for consistency
     of performance.

Here the first two steps of the cycle represent the design
phase of the problem solving strategy.  The last two steps of
the cycle represent the implementation and execution phases
respectively of the strategy developed.  These later processes
usually consist of translating the carefully designed problem
solution into a target software package.

An expert system is a computerised problem solving machine. It
can answer questions and act as an adviser on one of  a broad
range of subjects.  Expert systems are designed using simple
models which emulate the ways in which human experts solve
problems and are applicable in an ever broadening range of
subject domains. 

Computerised expert systems apply rules to solve problems. The
rules are either logical rules or rules of experience
(heuristics) gained from a human expert. These rules are stored
as a knowledge base of clauses that is used by the inference
mechanism of the expert system software to determine and follow
possible pathways to arrive at a conclusion.A major part of the development of an expert system is to
research the content of the topic and to organise the knowledge
gained as a series of facts which can be linked to the
conclusions by such rules. This aspect of the problem solving
process, called knowledge acquisition, will usually involve
consulting with one or more human experts, as authorities, in
the area under consideration and is mirrored in the first two
steps of the software development cycle. 

To implement the problem solution an expert system shell such
as EShell, Micro-Expert or VP Expert can be used. The advantage
of using an expert system shell is that it frees the student
problem solver from learning extensive language syntax to
concentrate more fully upon the semantics of the problem
solution. 

Alternatively, if formal programming skills are considered
important, then a computer language such as Prolog can be used
to implement the problem solution as Step 3 of the software
development cycle. 

Regardless of which software implementation strategy is
utilised, a study of expert system applications affords an
opportunity to provide materials which address the generalised
concepts, skills and problem solving strategies applicable to
students undertaking a secondary course in information
technology or other disciplines.

The process of creating the expert system provides students
with practice in determining and evaluating logical
relationships that occur in the subject domain under
investigation, in building models of problem solutions and in
producing complete reports. 

The process of creating an expert system can be further used to
complement any curriculum area for the teaching of cognitive
skills. For example, Science students could use an expert
system shell to produce a system to identify the chemical
elements. To complete this task they need to become actively
involved with the deep structure of Scientific knowledge, as
distinct from learning by rote the superficial symbol systems
of Science. 

Within secondary information technology courses and cross-
curricula subject areas, the problem domain may concentrate
upon expert systems which provide for advice, classification
and diagnostic applications.  Here the expert system determines
the facts during a consultation session with a user and zeros
in on a conclusion (backward chaining).Recent reductions in hardware costs, increasing processing
power and new software developments now mean that expert
systems are cost effective mechanisms with exploding interest
and educational applications. Schools and other educational
institutions are increasingly responding to such trends, and
the 'Technological Revolution' in general, by providing courses
of study which incorporate newer studies such as expert
systems.

A study of expert system applications provides opportunities
for delivering courseware which  

    is appropriate to secondary students;
    is teachable;
    is constant and based upon a proven methodology; 
    encompass objectives that are attainable by all students
     regardless of their background and experience;
    provides for a variety of cross-curricula learning
     experiences;
    provides for a variety of computer skills ranging from
     user friendly  shells through formal programming concepts;
    encompass sets of skills, concepts, generalities, logic
     and problem solving abilities which are of a continuing
     nature;
    provides a 'high-tech' unit of work;
    is interesting and fun.

The conference paper and session will focus upon 

1.   outlining possible approaches using software shells such
     as VP-Expert, Micro-Expert and EShell,

2.   outlining possible approaches using Prolog,

3.   outlining suitable project criteria, including assessment
     mechanisms, and 

4.   illustrating these using appropriate case studies.


References

1.   An Introduction to Expert Systems
     I J Lynch and J Komarmy
     Lynformation, 1993  

2.   Information Processing and Technology
     I. J. Lynch and J. Komarmy
     Morton Bay Publishing, 1991.

3.   Information Processing and Technology
     K. Offer and C. Thompson
     Jacaranda Press, 1991.

4.   Computing in Queensland
     I J Lynch
     Australian Educational Computing (Vol 5 Number 2) 

5.   Syllabus for IPT
     Board of Senior Secondary Studies, Queensland

6.   Practical Computer Methods Syllabus
     Department of Education, Queensland, 1988

7.   Turbo Prolog, Owner's Handbook
     Borland International, 1986

8.   Apprentice Computer Programmers
     B Harvey and J Richardson
     Australian Educational Computing (Vol 7 No 1)

9.   Information Systems and Expert System Shells
     T Managan 
     COMM 3 Journal, 1991

10.  EShell
     K Jacobi
     QSITE-IPT Teachers Conference, 1993

11.  I J Lynch
     A Case Study in Prolog
     QSITE-IPT Teachers Conference, 1993

12.  How To Solve It.
     G Polya
     Princeton University Press, 1947

13.  T Dwyer
     EShell Manual, 1992

14.  VP Expert (Manual)
     Software for the IBM PC
     WordTech, 1989

15.  Micro Expert (Manual)
     B Thompson and W Thompson
     McGraw-Hill, 1985

16.  An Introduction to Prolog.
     I J  Lynch
     Lynformation,  1994

17.  Introducing Expert Systems in Prolog
     I J  Lynch
     Lynformation,  1994