The fuzzy expert system is a form of problem solving used by a computer system, often used in the creation of artificial intelligence. Expert systems are types of decision-making computer software based on Boolean logic, meaning that the system uses a series of yes or no answers to try and solve a problem. Fuzzy expert systems expand on the traditional expert system and are based in fuzzy logic instead of Boolean logic. Fuzzy logic, as the name implies, means the answer is not a clear yes or no. It falls somewhere in the middle, and the computer must try to calculate an answer based on answers that may not be completely true but may not be entirely false either.
Known as the "father of fuzzy logic," Dr. Lotfi Zadeh introduced the concept of fuzzy logic in the 1960s while employed at the University of California in Berkeley. He published a paper in 1965 covering fuzzy sets. He explained not only the idea of fuzzy sets and logic, but also a framework for incorporating this new logic into the world of engineering. He also coined the term "fuzzy," in reference to this particular logical style, and the name stuck.
To understand the theory behind fuzzy expert systems, it is necessary to understand the basic concepts of Boolean logic and fuzzy logic. Though both rely on advanced mathematical algorithms, the core concept is simple. Both use answers to a series of questions or statements to formulate a new answer. In Boolean logic, the answers are either true or false, while in fuzzy logic an answer can be true, partially true, false, partially false, and several values in between, depending on what terms the programmer inputs into the program.
For example, if an expert system wanted to make a decision using Boolean logic, it would ultimately answer true or false, also referred to as yes or no. An expert system using fuzzy logic, however, could answer yes, no, maybe, or some other combination. It does this by drawing conclusions from its current knowledge base of information.
Knowledge bases are the heart of fuzzy expert systems. If a computer cannot come up with the correct answer, it is assumed the knowledge base does not contain enough information rather than assuming the program itself is wrong. The knowledge base might contain a statement such as "When x=yes and y=no then z=maybe." From this statement, fuzzy expert systems can conclude that when "x=yes" and "y=yes" that "z" must equal "yes" also, or that when "x=no" and "y=yes" that "z" still equals "maybe." If that is not the answer the programmer wanted, it means the knowledge base needs more information to come up with the correct answer.
Fuzzy expert systems make these calculations based on mathematical values. "Yes," "no," and "maybe" are assigned certain values. The computer looks at what the values of the terms in statements such as "x=yes and y=no" equal and adds their values. It then adds in any other relevant values and matches the final value with an answer like "maybe," "yes," or "no." Thus adding the mathematical values of "x=no" and "y=yes" tells the computer that the mathematical value for "z" equals "maybe."
