Expert Systems attempt to emulate the decision making abilities of a human expert using knowledge (facts) and inference procedures (rules). Perhaps the easiest artificial intelligence technique to master, it has also been arguably the most successful. Applications for expert systems run the gauntlet of computerization from medicine and engineering to games and business rules.
The Inference Engine Component Suite allows you to add an expert system to your application. Specifically, the TInferenceEngine component provides the engine for reading in your expert system and inferring new facts. You can enter the domain expertise for your application at design time using property editors (which are also available to you to be included in your run-time), at run-time using code, and at run-time by supplying text files in CLIPS format.
To build an expert system, you need to perform the following steps:
•Partition the knowledge in your expert system into multiple modules (optional)
•Define the types of objects in your domain
•Define the rules that infer new information based on your domain
•Define the initial facts about the domain
Optionally, you may need to also define:
Functions which can act on objects to extract information or modify the objects.
There are also a few tips to writing expert systems.
For more information, buy a book on expert systems or CLIPS such as "Expert Systems, Principles and Programming" by Giarratano & Riley.
