Flexible jobshop scheduling based on hybrid estimation of distribution and ant colony algorithm
                                                                                LU Honghao, LU Yujun
                     （Faculty of Mechanical Engineering & Automation, Zhejiang SciTech University, Hangzhou 310018, China)

Abstract: Aiming at the problem of flexible jobshop scheduling, an optimized method was studied, a function model for multiobjective flexible job shop scheduling problem was established, and a hybrid estimation of distribution and ant colony algorithm was given. It started with the estimation of distribution algorithm to quickly obtain the global superior solutions, then the pheromone initialization of the ant colony algorithm was improved by partly selecting the general superior solutions. Finally, the global optimal solution was quickly found by using the positive feedback mechanism of ant colony algorithm. In the improvement of this estimation of distribution algorithm, several methods to initialize machine selection and process sequencing were combined, and the corresponding probability model and population renewal method were given. Besides, in the improvement of this ant colony algorithm, the state transfer rule was described by establishing two path node sets, and the pheromone renewal mechanism was updated in phases locally and globally, which is beneficial to the rapid convergence of ant colony algorithm to the global optimal solution. The simulation analysis and comparison with other algorithms were carried out through two examples of flexible job shop scheduling. The results indicate that the hybrid estimation of distribution and ant colony algorithm has a good optimization effect and high efficiency in solving flexible job shop scheduling problems.

Key words: flexible jobshop scheduling problem（FJSP）; multiobjective; hybrid estimation of distribution and ant colony algorithm; positive feedback mechanism; global optimal solution