These are the two main methodologies used to develop expert systems (Efraim Turban 1999):

1. Large Scale Expert System SDLC.
2. Small Scale Expert System SDLC.

Large Scale Expert System SDLC is used for large size and safety critical projects. Safety critical projects involved cost and life risks. Small Scale Expert System SDLC is used for small size projects that are not safety critical.

Large Scale Expert System SDLC

Modified waterfall Software Development Life Cycle (SDLC) is called as Large Scale Expert System SDLC. These are the main phases of Large Scale Expert System SDLC:

• Project Identification & Selection.
• Project Initialization.
• System Analysis and Design.
• Rapid Prototyping.
• System Development.
• Implementation.
• Post Implementation.

Project Identification & Selection

There are so many projects in every organization. The analyst or developer defined the main projects and these are prioritized based on few factors:

• Urgency.
• How much a project will add cost and value to organization.
• Expected budget.
• Expected schedule.
• Level of technical difficulties or constraints.

CEO/Board of Directors/Board of Governors selects the Expert System (ES) project to be developed according to the need of organization. The developer should work on these factors before developing an expert system:

• Clear problem definition.
• Development Strategy.
• Development Methodology or Process.
• Budget.
• Schedule.
• Working Location.

Knowledge Engineer should have clear understanding before developing an expert system. Selection of development strategy should be finalized between developer and customer, such as development of customized software or purchasing of general purpose software. Selection of Large Scale Expert System SDLC or Small Scale Expert System SDLC if customized software strategy is adopted. Budget and Schedule are risk factors for the expert system development because there is no guarantee for its budget and schedule. Knowledge Engineer must consider the location where Expert System (ES) is going to be implemented.

Project Initialization

These are the main phases of project initialization:

• Problem or Project Definition.
• Assessment of Needs.
• Generating Alternative Solution.
• Verify Approach to Develop Expert Systems.
• Management Issues.

Problem or Project Definition & Assessment of Needs

Knowledge Engineer works on the problem to define it and tries to find the source of the problem. Problems could be:

• Lack of productivity.
• Poor quality.
• Overtime charges.
• Lack of technical expertise.
• Lack of human expert.

Generating Alternative Solution

He developed alternative solutions of the problem/s if lack of knowledge is the main cause behind all those problems mentioned above. The alternative solutions could be:

• Arrange training sessions to equip the workers with technical expertise.
• Arrange manuals for the technical people.
• Recommendation of hiring an expert.
• Buy off the shelf package.
• Develop conventional software instead of developing an expert system.

Verify Approach to Develop Expert Systems

These are the three main tasks to verify the expert system approach:

1. Requirements.
2. Justification.
3. Checklist.

Knowledge Engineer works on the basic requirements to justify and prepares a checklist before making a detailed analysis and design. Knowledge Engineer performs these activities to verify the basic requirements:

• Identifies tasks that need common sense.
• Identifies tasks that do not need common sense.
• Defines and understands all tasks.
• A contract is made with an expert to complete the Expert System (ES).
• Test data is prepared.

Expert system approach should be justified:

• Will the solution generate large amount of money after implementation of Expert System (ES)?

• Will the system preserve the knowledge of an Expert?

• Will the Expert System (ES) operate at remote locations where human experts are not available?

• Will the Expert System (ES) work in a hazard environment where life risks are involved?

Expert System (ES) development is the best choice if the answer is yes for any question or questions mentioned above. Knowledge Engineer prepares a checklist before developing an Expert system (ES):

• Degree of accuracy level required making decisions by the Expert system (ES).

• He must have complete set of examples about the solution.

• End users of ES should be able to judge the wrong decisions made by the Expert system (ES).

Management Issues

Knowledge Engineer has to look upon these management issues:

• Who is going to finance the project?
• Who is or are the suspected customer or customers?
• Are the required resources available to develop the project?
• Who is in favor of expert system’s project from the top management?

Related posts:

1. Methodologies to Develop Expert Systems (Part Three) Rapid Prototyping Rapid Prototyping is divided into these main sub...
2. Methodologies to Develop Expert Systems (Part Two) System Analysis and Design Main phases of System Analysis and...
3. Main Purposes of Developing Expert Systems and Main Concepts of an Expert System (Part One) Main Purposes of Developing Expert Systems The main purposes of...
4. Components of an Expert System These are the main components of an expert system: Knowledge...
5. Main Areas of ES-Advantages and Disadvantages of ES-Comparison of Conventional vs. Expert System Main Areas of Expert Systems Expert System is always developed...