This approach to the world easily leant itself to digital logic and the early applications of computer 

programming. 

	However, this classical approach has its limitations. The real world has many situations where 

belonging or not belonging to a group has degrees of truth. As an example, consider my twelve year old 

daughter. To us adults, she is very young, but to her 13 year old sister, she is not very young at all. So 

she both belongs to and is excluded from the set of “very young”. 

 	This problem was described as early as Aristotle and Plato, and later by Hegel (Philosopher 

and Logician) and others. In the early 1900's, Lukasiewicz described a three-valued logic, along with the 

mathematics to accompany it. This third value was “possible” and belonged somewhere between true and 

false. Lukasiewicz later theorized an infinite number of values between true and false.

	To resolve the issue of degree of belonging, Lotif Zadeh developed Fuzzy Logic in 1965, in 

which any value between 0 and 1 can be assigned to indicate the degree of belonging. When Lotif Zadeh 

first published his research on Fuzzy Logic in 1965 he was criticized by his U.S. colleagues. However, the 

European and Asian math and science communities explored this “new” philosophy as a continuation of 

Lukasiewicz's work. 

	The United States was heavily invested in “Expert-Knowledge” based systems and was 

reluctant to abandon their research in favor of the much more radical and simpler to implement Fuzzy 

Logic based inference system. 

	This inflexibility on the part of the United States mathematics and science communities to 

augment their binary logic systems allowed the Asian and European communities to employ Fuzzy Logic 

controls applications without U.S.competition.

History of Controls Systems

	Throughout the history of control automation mankind has tried to mimicthe human worker 

to control machines and processes without human interaction. Punch cards to control Jacquard looms 

and the fly-ball governor to control steam engine speed are just two examples of “hard” control systems. 

These systems are limited by the fact that they are not easily modified and are incapable of learning. 

	The emphasis in mechanical control systems gave way to electrical control systems since 

WWII. Many research projects focused on how to make computers mimic humans in the acquisition, 

storage and logical use of information. In short, how to make a computer control system that can replace 

a human expert. 

	This research field is called artificial intelligence. Early work emphasized knowledge-based 

expert systems. These systems, such as the 1976 medical diagnosis system, MYCIN, relied upon a series 

of shells that successively narrowed down options until a single correct was found. This technique was best 

employed in static situations such as the diagnosis of medical patients. It required enormous amounts of 

storage and was slow implement.

	Therefore these massive expert-knowledge-based systems were too slow to control dynamic 

systems. So industry emphasis centered on ladder logic systems first utilizing relays and later using PLC’s. 

The limitations with these control systems are the lack of flexibility and an inability to react or learn when 

exposed to new situations. It was when overseas competitors began successfully employing Fuzzy Logic 

in consumer electronics on a wide scale that U.S. manufacturers and researchers began to explore the 

flexibility of Fuzzy Logic In the 1990s research expanded into other options in Artificial Intelligence, 

including neural networks and fuzzy control. In general neural networks are more effective in areas requiring 

data analysis and fuzzy logic is better for control. This has given rise to the hybrid of the two systems called 

fuzzy-neural networks. 

HOME