Abstract

The dynamic facility layout problem (DFLP) deals with the arrangement of facilities/departments in a factory for different periods so that the location of the facilities can be changed from one period to another one. Traditionally, this problem is formulated to minimize the sum of material handling and rearrangement costs in the planning horizon by assuming that all parameters are deterministic. In this paper, we assume that the material flow between departments and rearrangement costs are uncertain and, accordingly, develop the robust counterpart (RC) of the DFLP model. The model is computationally intractable; therefore, we propose a hybrid algorithm based on nested partitions (NP) and simulated annealing (SA) algorithms, namely NP-SA. Moreover, we develop a heuristic algorithm to compute the values of the additional variables used in the RC model. The numerical results indicate that the NP-SA algorithm is very effective in giving a good solution in a short time. Furthermore, a simulation study demonstrates that, on average, robust solutions are better than nominal solutions.

摘要

动态设施布局问题（DFLP）处理工厂中不同时期的设施/部门的布置，以便可以将设施的位置从一个时期更改为另一个时期。传统上，通过假设所有参数都是确定性的，制定此问题的目的是将物料搬运和重新安排成本的总和最小化。在本文中，我们假设部门之间的物料流和重新安排成本是不确定的，因此，我们开发了DFLP模型的鲁棒对应物（RC）。该模型在计算上难以处理；因此，我们提出了一种基于嵌套分区（NP）和模拟退火（SA）算法的混合算法，即NP-SA。此外，我们开发了一种启发式算法来计算RC模型中使用的其他变量的值。数值结果表明，NP-SA算法在短时间内给出良好的解非常有效。此外，仿真研究表明，平均而言，健壮的解决方案优于名义解决方案。