Genetic algorithms in learning fuzzy hierarchical control of distributed parameter systems

A GA-optimized, two-level, fuzzy hierarchical controller which utilizes both spatial and temporal measured data for control of distributed parameter systems is discussed in this paper. A distributed model of a single uniformly flexible link illustrates a possible control configuration. A Dynamic Fuzzy package is developed in C for interfacing with the simple genetic algorithm (SGA) which is a C translation of Goldberg´s SGA in Pascal. Dynamic Fuzzy allows for creating new rules, disposing of old undesirable rules and tuning of the membership functions. Dynamic Fuzzy´s memory management and data handling is also designed to reduce memory access time for real-time applications. Specifically, the real time hardware aspects of such a hybrid control is discussed using the Texas Instrument´s TMS320C30 DSP chip and the DSP Research´s TIGER 30 board