Study on multi-objective flexible production scheduling based on improved immune algorithm
Yu Jian-jun,Xu Xu-jun,Fei Ye
DOI: https://doi.org/10.1109/ICMSE.2008.4668967
2008-01-01
Abstract:Because of multi objective, multi-constraint, dynamic flexible process route and complex modeling, multi-objective flexible Job Shop scheduling (MOFJS) is more complicated than the classical Job Shop scheduling, and therefore it can't generally been solved by the ordinary optimization methods. Firstly, the model of MOFJS is set up. Secondly, immune algorithm based on excellence holding is put forward. In the algorithm, the best antibodies are held and utilized by immunity memory and the local best genes are held and utilized by immunity vaccine in every alternation to accelerate algorithm convergence. Thirdly, in view of the flexible process route of the scheduling, a novel double layer antibody coding method based on operation and machine is brought forward, and an effective decoding method based on machine capability space is developed. In addition, the multi objective ranking evaluation technique is adopted to evaluate and optimize several contradiction objectives, such as time, equipment and cost. Finally, the availability and the superiority of the algorithm, the strategy and the model are validated by the simulation of Benchmark Job Shop scheduling problems and the scheduling instance of Xi'an Aero-engine (Group) LTD. in China. © 2008 IEEE.