Genetic algorithms for modelling, design, and process control
Charles L. Karr · 1993
Genetic algorithms (GAs) are becoming increasingly popular as tools for solving search, optimization, and machine learning problems.The true strength of these search techniques lies in their ability to perform impressively across a broad spectrum of problems; they are robust.From system modelling to engineering design to process control, GAs have been used effectively to solve problems that have given more conventional search schemes considerable difficulty.This paper provides an overview of three projects involving GAs that have recently been completed by researchers at the U.S. Bureau of Mines, Tuscaloosa Research Center.The three projects are: (1) using a GA to solve a modelling problem using the Ree-Eyring equation, (2) designing a piece of separation equipment an air injected hydrocyclone, with a GA, and (3) tuning a fizzy logic controller with a GA.These three applications provide an indication of the potential of GAs, and represent "state-of-the-art" of GA use in the minerals industry.