基于Bayesian统计推理的分布估计算法求解, 柔性作业车间调度问题

在Bayesian统计推理理论的基础上, 提出一种新的求解柔性车间调度问题的分布估计算法.首先, 根据所有工件的工序排列顺序提取进化过程中种群的优良信息, 建立一个不断更新的先验分布概率模型, 再以相邻工序出现的频率为基础建立条件概率模型; 然后, 结合两个模型的信息使用Bayesian公式建立一个后验概率模型, 该模型综合了进化过程中不断更新的优良信息和相邻工序出现的频率信息, 可用以更好地指导产生新群体.仿真结果表明算法具有较好的寻优能力.

Solving flexible job-shop scheduling problems with Bayesian statistical inference-based estimation of distribution algorithm

A new estimation of distribution algorithm for flexible job-shop scheduling problems based on Bayesian statistical inference theory is proposed.At first,according to the permutation of all operations, the model of priori distribution probability is built extracting the information of superior solutions updating, then the model of conditional probability is also built based on the frequencies of neighbor operations appearing.After then,the model of posterior probability is given by combining the above two models to guide new population generating with Bayesian formula.Such model is characteristic of guiding now population generating well,for it synthesizes the information both of updating knowledge and of neighbor operations appearing frequencies.The simulation results show that the proposed algorithm has the preferable search ability.