CASIA OpenIR  > 毕业生  > 博士学位论文
钢铁企业CIMS下生产计划的研究
其他题名Study on Production Planning In lron & Steel Enterprises CIMS
蒋圣平
学位类型工学博士
导师邹益仁
2004-04-01
学位授予单位中国科学院研究生院
学位授予地点中国科学院自动化研究所
学位专业控制理论与控制工程
关键词年度生产计划 批量组合计划 决策支持系统 模型 Agent 多agent系统 遗传算法 启发式算法 Annual Production Planning Batch Planning Dccision Support System Model Agent Multi-agent System Genetic Algorithm Heuristic Alg
摘要本文以钢铁企业CTMS的生产计划问题为研究背景,针对其中涉及的一些 年度生产计划、炼钢-铸批量组合计划及决策支持问题,展开了一系列模型 与算法的研究。进行此类问题的研究,不仅具有实际的应用价值,同时也丰富 了生产和管理的研究内容。 论文的主要内容和创新之处如下: 1.模型自动生成技术研究。基于传统的顺序建模系统,提出了一种基于 多Agent的模型自动生成方法,该方法结合了面向对象和人工智能技术,运用 大系统分解协调理论,构造了一个并行的、自主的、可扩展的、具有良好适应 性的建模系统。论文详细介绍了该系统的体系结构、设计方案及关键技术,并 通过年度生产计划模型生成进行了案例说明。 2.钢铁企业年度:生产计划的研究。结合钢铁企业的生产工艺和年度牛产 计划的实际需求,建立多分厂、多机型、具有能力约束的年度生产计划 模型。该模型是一个多层递阶优化问题,对此,论文提出了一种基于模型转 换、并将遗传算法与二段法相结合对模型进行求解的方法。仿真实验表明了 该模型和算法是可行的。 3.炼钢-铸批量组合计划的研究。炼钢-铸批量组合计划分为最优炉 次计划和最优浇次计划。论文建立了一个考虑合同是否可分的炼钢-铸最优 炉次计划数学模型,提出了两种改进的启发式遗传算法,并对遗传算法的编 码、解码及算子操作等做了研究,仿真实验验证了孩模型与算法是有效可行 的。运用不同的启发式规则,将启发式遗传算法应用到无剩余炉次的最优浇次 计划问题中,也取得了很好的效果。 4.年度生产计划决策支持系统的研究。基于传统的模型驱动的决策支持 系统,结合Web技术,提出了一种基于Web的年度生产计划决策支持系统体 系结构。论文详细描述了模型驱动的决策支持系统的设计方案和主要内容,并 以此为基础开发了钢铁企业年度生产计划决策支持系统软件平台
其他摘要Based on the Production Planning for Iron & Steel Enterprises, this dissertation does some researches on Decision Support System (DSS) and several optimal models and algorithms of Annual Production Planning (APP), Batch Planning (BP) of steel making and continuous casting. These researches are practically significant and enlarge production planning and management fields. Main works and contributions included in this dissertation are described as follows Firstly, this dissertation presents an agent-based approach for model construction combining an object-oriented technology with artificial intelligence. The approach has the advantages of not only inheritance and encapsulation of object-oriented but also reasoning study of artificial intelligence. Based on it, the conventional sequential system for model construction is remodeled into a concurrent and autonomous and scalable one that can adapt to the changes of the environment. In this paper, the architecture and local representation of an agent and the key technology how to design an agent-based system are described in detail, and an example for APP is presented to illustrate the usefulness of this approach. Secondly, according to the work procedures of Iron & Steel enterprises and the actual requirement of APP, a related math model is proposed with multi-plants and multi- machines and capacity constraints. The model belongs to a multi-level hierarchical optimal problem and can't be solved easily at present, so this dissertation introduces a new idea of transferring; this multi-level model into a bi-level one according to the structure of the products. Based on the model transformation, this bi-level model is calculated by combining genetic algorithm with two-phase method. Based on model transformation, the problem is solved by combining genetic algorithm with two-phase method. The simulation results show that this model and its algorithm are feasible. Thirdly, BP of steel making and continuous casting consists of Charge Design Problem (CDP) and Cast Plan (CP). In this dissertation, a new CDP model is developed to improve the efficiency of charges and two related Heuristic Genetic Algorithms (HGA) are proposed. HGAs combine different heuristic rules with genetic algorithm and adopt domain-specific encode scheme, heuristic decode procedure and proper genetic operation to get more correct solution at higher speed. The experiment results prove that the model and its algorithms are valid and feasible. HGA with different heuristic rules is appropriate to CP. Finally, this dissertatiobn proposes a web-based DSS architecture by combining the conventional model-driwn DSS with Web technology. Moreover, the key issues and rules are described in detail to implement a model-driven DSS and the APP software is developed for Iron & Steel enterprises.
馆藏号XWLW805
其他标识符805
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/5794
专题毕业生_博士学位论文
推荐引用方式
GB/T 7714
蒋圣平. 钢铁企业CIMS下生产计划的研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2004.
条目包含的文件
条目无相关文件。
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[蒋圣平]的文章
百度学术
百度学术中相似的文章
[蒋圣平]的文章
必应学术
必应学术中相似的文章
[蒋圣平]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。