A method to reduce fuzzy control rule sets
K. Komiya, K. Thuneishi, Kenji KUROSU, K. Kagekawa · 2003
The number of fuzzy control rules increases rapidly along with the increase of input variables to the fuzzy control. It is very hard for human experts to set a large number of rules necessarily to control the system. We propose a method to reduce the rule sets by introducing pseudo variables. With the reduced fuzzy control rule set we can control systems to goal states almost the same as with the original control set. Using reduced fuzzy rule sets, we carried out computer simulations of pendulum systems, one is to control a single pendulum standing still and the another is to control a compound pendulum (two weights) to reduce the upper weight's swinging with the first standing on the base.