多目标规划在储药柜设计中的应用

Application of Multi- Objective Optimization( MOO) in Medicine Storage Design

在设计储药柜最优化方案时将部分限制条件进行预先处理,从而降低0-1变量个数,引用0-1线性规划模型。在此基础上将问题转化为多目标规划模型,进而采用LINGO软件求解得到有5种槽宽,此模型及软件使优化问题得以合理解决,但与实际的差别较大。进一步利用聚类分析对槽高类别进行优化,然后通过聚类结果的比对得到答案槽高为7种,此方法能够直观看出各变量的关系,且结果精确。

In this paper we focus on the optimization design strategy of Medicine Storage. The 0 - 1 linear program- ming model is incorporated, and some constrained are pre - processed to reduce the number of binary variables. The problem is reformulated as a multi - objective optimization （MOO） model. It turns out that the optimal result is 5 distinct slot widths are needed, which is a reasonable answer to the model but not practical. The improved model uses clustering analysis to optimize the number of distinct slot widths, and obtained the more accurate result of 7 distinct slot widths, which could directly reflect the relationships between the variables.