Abstract: According to the characteristics of flatness error evaluation, a novel improved genetic algorithms based on real coding is proposed. A generation-alternation modei based on the minimal generation gap (MGP) and blend crossover operators (BLX-α) are employed. Compared with conventional methods, it has the advantages of simple algorithms, strong robustness and high op timization efficiency. In the meantime, the objective function calculation approach in flatness error evaluation by genetic algorithms is presented. Finally, the experiment results evaluated by differ-ent methods show that the improved genetic algorithm converges to the global optimum more rapidly than conventional methods.

Key words: genetic algorithm, reai coding, flatness error