Non-Fuzzy Knowledge-Rule-Based Controllers and their Optimisation by Means of Genetic Algorithms
Ivan Sekaj · 2000
A non-fuzzy rule-based system for process control and general applications is described. The system uses a similar rule-base as fuzzy systems, but it does not use fuzzy variables. The items of the knowledge-base are only real numbers and the evaluation mechanism is based on a n-dimensional interpolation. It is very simple and computationally fast. Next a genetic algorithm-based optimisation of the rule-base is shown. The proposed approach is demonstrated on process control simulations and real-time examples as well.