Most scheduling problems are highly complex combinatorial problems, However
, stochastic methods such us genetic algorithm yield good solutions.
In this paper, we present a controlled genetic algorithm (CGA) based on fuz
zy logic and belief functions to solve job-shop scheduling problems.
For better performance, we propose an efficient representational scheme, he
uristic rules for creating the initial population, and a new methodology fo
r mixing and computing genetic operator probabilities.
A 10-jobs/6-machines example shows the effectiveness of the developed metho
d.