文档介绍：华中科技大学硕士学位论文
摘要
随着我国市场经济的不断健全和发展,烟草企业的生产方式正逐步从过去的完
全由国家计划转为由市场导向,即按客户订单和销售合同来组织货源、安排生产。
企业生产计划不但要以市场需求来确定,还要根据企业制造资源的实际能力和库存
来调整。因此设计行之有效的自动排程系统,对于提高企业生产效率,增强市场竞
争力有着重要的意义。
鉴于实际生产排程优化问题的复杂性、约束性、非线性、多目标、建模困难等
特点,寻求合适的建模与优化方法是我们要研究的主要问题。因此本文采取了层次
化建模方法,将整个生产计划制定流程分为由上至下的两个层次:需求分配层和生
产点排程计划层。分别确定了这两个层次的优化目标和约束条件,继而构建了数学
模型。在构建生产点的排程计划模型时,我们又将整个生产流程分为制丝和卷接两
个阶段来考虑,分别建立了数学模型进行优化。
生产排程问题通常具有 NP 困难的特点和非线性的特点,因此我们采用了遗传算
法来求解模型的优化结果。在需求分配层中,我们针对问题规模相对较大的特点,
采用了基于生成树的遗传算法来求解;而各生产点的排程问题中的卷接优化模型具
有多目标的特点,我们采用了适应性权重的方法,构造了权重和目标函数,来解决
这一问题。
我们根据一个实际的算例,使用 编程求解,得到了较为理想的优化结
果,我们将这一排程结果与武烟集团的手工排产结果进行比较,发现自动排程结果
在降低库存成本,提高生产的连续性方面具有优势。
关键词:生产排程,遗传算法,生成树,多目标优化
I
华中科技大学硕士学位论文
Abstract
With the continuous development of China’s market economy, the production
planning of o enterprises which used to be controlled by government, is now
transforming to market-oriented mode gradually, which means, the enterprise must
schedule production plan according to contract needs. Moreover, enterprises should
analyze the capabilities of production line and inventory, in order to arrange and adjust
production plan properly. So it is meaningful to design an automatic system of Production
Scheduling to help enterprises to improve productivity and strengthen their core
competitiveness.
Production planning problems are plex nonlinear multi-objective
programming problems with constraints. In consideration of the natural characteristic of
practical production planning problems, we use hierarchical approach to analyze the
process of production planning. We divide the model into two levels. The objective of the
higher level is to distribute the aggregate demand among each factory properly, while The
task of the lower level is to arrange the production schedules of every factory accurately
and efficiently. The detail of each level is enriched, and an appropriate algorithm is
developed and applied.
To solve the NP-hard pr