Development of Expert Systems

Development of Expert Systems:

There are basically following steps that are involved in developing any expert systems:

1. Identify the problem: Like many compiler programs the expert systems are in a sense a solution looking for a key problem. To justify the development of an expert system there must be a real problem in order to solve. For this particular reason, the first step in the development of an expert system should be to examine the situation and then clearly decide what the problem is and how much the system may be helpful.
   
   
2. Consider the alternatives: Although the problems may fit the criteria for an expert system we should be careful with simpler or equally satisfactory alternative solutions. For example, certain kind of employee's performance can be connected to training. However, the solution may be to provide the employees with the information they need in a written manual. A solution may not be the best but it should be at least the simplest and least expensive.
3. Feasibility: The next step is to determine whether the development system is feasible or not. The system should be feasible from the every aspect that is technical, economically and so on.
4. Selection of development tools: An expert system development tool is a software package that allows us to enter the expert's knowledge into the computer without having to program. Most of the expert system development tools are rule based. However, some tools permit the implementation of the frames and semantic network but they are little expensive.
5. Perform the knowledge engineering: At this point in the development process we are finally ready to perform some real creative work. The development of an expert system actually begins with the knowledge engineering that is acquiring the knowledge which comes from different fields like a standard textbook, article or other references.
6. System design: As we have selected the appropriate tools that are required using the knowledge we can now begin to design the expert system. First, we need to create an outline for a hierarchical flowchart, matrix decision tree or other formats that will help us for organizing and understanding the knowledge. Using these aids we will be able to convert the knowledge into the "if-then" rule. Once the basic design is complete we can begin to create a prototype of one of the segments of the system."
7. Complete development: Once we are satisfied that the system is going to work properly we can begin to expand the prototype into the final system.
8. Testing and debugging: After the expert system has been developed we need to spend more time for the testing purpose. There will not be any such expert system which is perfect for the first time and a considerable amount of work will be required to validate it. The feedbacks of the users will show the places to make the corrections so that we can achieve the best performance.
9. Maintenance: The important part of the expert system development is the ongoing maintenance, updating the system with new knowledge and removing the knowledge that is no longer applicable.

Limitations of expert system:

• It is difficult to maintain.
• It is not widely used.
• It is found to be limited to the relatively narrower domain.
• There is a possibility for errors.

References:

1. Elaine Rich, Kevin Knight 1991, "Artificial Intelligence".
2. Nilsson, Nils J. Principles of Artificial Intelligence, Narosa Publishing House New Delhi, 1998.
3. Norvig, Peter & Russel, Stuart Artificial Intelligence: A modern Approach, Prentice Hall, NJ, 1995
4. Patterson, Dan W. Introduction to Artificial Intelligence and Expert Systems, Prentice Hall of India Private Limited New Delhi, 1998.