Abstract: A fast genetic algorithm——GSAGA(generalized self-adaptive genetic algorithm) is presented in this paper First,evenly distributed initial population is generated. Then, high quality immigrants are introduced according to the condition ofthe population schema. Finally, crossover and mutation are operated on self-adaptively in GSAGA, the searching performance and global convergence are greatly improved compared with many existing genetic algorithms. Through emulation, the validity of this modified genetic algorithm is proved..