Optimization of adaptive control rule of dead time system by genetic algorithm

The authors propose a parallel processing genetic algorithm (PGA) with fuzzy reasoning for optimizing the adaptive control rules of dead time systems. The fuzzy reasoning is applied to adjust the population size of each operator because a specific operator may suit for a certain stage of search. It adjusts population size by sensing the accumulate increment of fitness values indices. Simulation results show that the PGA with fuzzy reasoning helps searching out an optimal solution better than the traditional methodology. On the other hand, the new added operator, sub-exchange, shows power in search and the search time is thus reduced.