An adaptive Genetic Algorithm for the Flexible Job-shop Scheduling Problem

Aiming at the stage-related characteristics in solving process of the Flexible Job-shop Scheduling Problem (FJSP) and the evolution characteristics of Genetic Algorithm (GA), an Adaptive Genetic Algorithm (AGA) is presented in this paper, combined with existing GA problems for solving FJSP. In the solving process, AGA automatically adjusts selection rate P_s, crossover rate P_c, mutation rate P_m and other operation parameters according to the genetic generation. Furthermore, it automatically changes scope of action of crossover operator and mutation operator on the chromosome. In the AGA solution scheme, linear interpolation method is used to realize automatic change of operation parameters from initial stage to middle stage; extended power function implements adaptive adjustment in the function scope of operators. Meanwhile, coding method is adjusted correspondingly for realizing operator adaptation. Instance simulation verifies that the FJSP´s own characteristics are utilized in its solution by using AGA, which overcomes traditional GA´s limitation. Such a method has a relatively high search power during the whole solution process, especially in the end of the process. And solving efficiency and precision are improved greatly.