Technology has always been about building better, faster, and smarter machines. Expert systems embrace this concept by using advanced computer logic to create software that appears to "think" and make decisions on its own. Traditionally built on Boolean logic — logic using only true or false values — expert systems use complex algorithms to calculate answers from a large database of information. If the computer cannot determine the correct answer, it is assumed not that the program is wrong but that the knowledge base does not contain enough information on the subject.
When a computer must make a decision, it all breaks down to a series of true or false statements. If programmed to light up when a button is pressed, then pressing the button sets it to true and not pressing the button sets it to false. False means no light while true turns the light on. This is the basis of computer logic.
An expert system takes these true and false answers to a new level. By combining a series of true and false answers, the computer tries to determine how to react to a certain situation. It may change its response based on the specific pattern and number of true and false answers.
The idea behind these systems is based on how people think. Humans can store vast amounts of new knowledge and make decisions based on previous knowledge. The computer is programmed to “think” and make decisions based on the knowledge found in its database and on its previous experiences. In a fashion, it’s as if the computer is "learning" from its past successes and failures.
Two main forms of expert systems exist. The traditional expert system uses Boolean logic to makes its decisions. A fuzzy logic expert system, on the other hand, does not. It calculates a range of values that fall in between simple true or false answers to determine to what degree a statement is more true or more false.
Fuzzy expert systems are more human-like than traditional expert systems in the way they "think." These expert systems are not told specific answers to a problem, but rather given one statement from which they draw additional conclusions. This process is known as inference.
For example, if a statement reads "All female cats are striped. Miss Kitty is a female cat," fuzzy expert systems would infer that since all female cats are striped and Miss Kitty is a female cat, then Miss Kitty must be striped. Fuzzy logic can also calculate more complicated values, such as determining the likelihood of a specific female cat being striped if only a percentage of female cats have stripes. Traditional expert systems would need much more instruction to reach these same conclusions.
