Neural networks for process control

Abstract

In this paper we present motivation for pursuing different classes of adaptive artif7ial neural network controller designs. A literature review of the current state of the art in artif-xial neural network controllers, and a framework for conducting research in the field is included. In Addition, hybrid neural nehvorklfuzzy controllers are described. The importance of benchmarking comparable adaptive eontrollers is discussed, and appropriate benchmarks are proposed. regulators (STR), adaptive inverse controllers (AIC) and &el referme adaptive controllers ( M U C ) , among others 121. h o t h e r way to classify these controllers, is by their use in a direct Of an indirect controller stnrcture as described byAlso there are applications of A"C,s which are not purely ANN but which include the use of linguistic or heuristic features, as used in Fuzzy Logic Controllers.