On applying information principles to fuzzy controls

It is shown that the uncertainty principle follows from certain natural assumptions about consistent decision making and serves as a basis for resolving the ambiguity of missing values and thus for deciding the precise nature of control rules.

Abstract

The use of fuzzy control techniques requires categorizing membership grades for various controlling (input) and controlled (output) parameters. In many engineering situations, such grades are specified, a priori, only partially; several values might either be missing or be presented as subject to certain constraints. The principle of maximum uncertainty, a fuzzy analogue of maximum entropy, would then serve as a basis for resolving the ambiguity of missing values and thus for deciding the precise nature of control rules. It is shown that the uncertainty principle follows from certain natural assumptions about consistent decision making.