Research on genetic algorithm strategy selection for function optimization based on function cluster

It is difficult to determine suitable strategy selection for function optimization using a genetic algorithm, so it is still an important problem on genetic algorithms research. It is suggested that the function be first clustered into different types using fuzzy C-means algorithm in accordance with the landscape information of function. Then, the solution schemes of strategy selection guidance knowledge should be given according to the types. This knowledge can guide the genetic algorithm to do the function optimization with the optimal strategy. Experimental results show that the method can reasonably determine the optimization strategy for a category of functions in optimizing a genetic algorithm. This research provides a new kind of method to obtain the optimal strategy for the optimization of a genetic algorithm. © 2011 ICIC International.