Some aspects of synthesis of probabilistic fuzzy controllers (Q800859): Difference between revisions

Described is an approach to synthesis of probabilistic fuzzy controllers. They are thought as control algorithms based on a finite number of statements (implications) of the form ''if \(A_ i\) and \(B_ k\) then \(C_{ik}''\), \(i=1,2,...,N\), \(k=1,2,...,M\). \(A_ i\), \(B_ k\), \(C_{ik}\) are viewed as probabilistic sets as introduced by \textit{K. Hirota} [ibid. 5, 31-36 (1981; Zbl 0442.60008)] and expressed by their distribution functions. For any input of the controller its output is calculated by means of sup-min composition. Compact formulas for the output are obtained by the use of the so-called intersection coefficient. Moreover, an idea of decomposition of multi-dimensional controllers is studied.