求解不等面积设施布局问题的交互式分布估计算法

An interactive estimation of distribution algorithm for unequal area facility layout problem

中文关键词  进化算法  交互  不等面积设施布局  分布估计算法  概率模型

英文关键词  evolutionary computation  interactive  unequal area facility layout problem  estimation of distribution algorithm  probability model

中文摘要

      同时考虑定量和定性指标, 采用交互式进化优化方法求解不等面积设施布局问题可以获得更好的鲁棒解.基于此, 本文提出一种双概率模型交互式分布估计算法. 首先, 统计群体信息, 构建显式指标概率模型, 估计决策变量分布. 其次, 基于决策变量表现型相似度, 构建隐式指标概率模型, 同时, 基于效用函数估计个体定性指标. 然后,将两个概率模型合并成双概率模型, 采样生成新种群. 最后, 基于推荐个体和用户评价信息, 动态更新两个概率模型. 将所提方法与6种相关进化优化算法对比, 在纸品处理车间问题和16个不等面积设施布局问题测试集上的运算结果表明, 所提方法可以高效获得最优布局方案.

英文摘要

      Both the quantitative and qualitative indices should be considered in order to obtain more robust solutions in the unequal area facility layout problem (UA-FLP) with interactive optimization method. This paper proposed a dual-probabilistic-model-assisted interactive estimation of distribution algorithm. Firstly, an explicit index probability model was established to estimate decision variables distribution through making an statistics to the group information. Subsequently, an implicit index probability model was established based on phenotype similarity of decision variables. In this way, the individual qualitative index was estimated based on utility function. Furthermore, the two probability models were merged into dual probabilistic model which generated new population through sampling. Finally, the dual probabilistic model was dynamic updated based on recommended individuals and evaluation information. The proposed method was compared with six related evolutionary algorithms on the Carton Packs problem and 16 UA-FLP test sets, and experimental results show that the proposed algorithm can efficiently obtain optimal layouts.