上海交通大学硕士学位论文 1热轧生产计划优化系统研究与应用 摘 要摘 要 生产计划和调度问题是伴随制造企业整个生产管理过程的最重要的问题之一，是衡量企业竞争力的重要指标。钢铁企业生产中的很多生产计划和调度问题，这些问题都可以归结为组合优化问题，如炼钢调度，连铸计划，热轧计划，物流平衡，运输调度，一体化生产调度等。组合优化问题是 NP-Hard 问题，而这类问题是无法使用传统的优化方法在有限时间内获得最优解的，必须借助启发式算法或其他智能算法。 热轧是钢铁企业三大关键工序炼钢、连铸、热轧之一。热轧计划编制是钢铁生产中一个重要环节，是一个组合优化问题。热轧计划的编制不仅关系到热轧生产，还关系到炼钢、连铸生产的进行。早期，计划员采用手工方式进行计划编制，由于人力的局限性，计划员通过手工方式来编制计划的速度比较慢，编制出的计划效果往往不是很好，甚至较差。而如今国际上先进的钢铁企业大都采用模型的方式进行计划编制，这种方式能同时编制出多个计划，且编制速度快，效果好。 文章研究了宝钢热轧计划编制的实际问题，分析了宝钢热轧计划的相关规程约束，建立了计划编制 MTSP 模型，并在此基础上给出了求解该问题的有效的智能优化算法遗传模拟退火算法。论文描述了宝钢自动化所在热轧计划优化编制问题上开发的应用系统，该系统成功地解决了热轧计划的自动编制问题，并取得了突出的效果： ? 提高了工作效率、减少了劳动强度 ? 提高了计划的合理性与科学性 ? 满足产量均衡，提高流量命中率，轧棍利用率 ? 提高热送热装率，从而大大降低成本，提高了效益 上海交通大学硕士学位论文 2关键词关键词： 组合优化、建模与优化、炼钢、连铸、热轧、热轧计划、遗传算法、模拟退火算法、旅行商问题、应用设计 上海交通大学硕士学位论文 1Research of Hot Rolling Planning Optimization & Application System Abstract Production Planning and Scheduling, which evaluate the competence of the company, is one of the most important problems in Manufacturing Enterprise. There are many Planning and Scheduling problems in Steel Manufacturing Enterprise, such as Steel-making Scheduling, Continuous-Casting Planning, Hot Rolling Planning, Logistics Balancing, Transport Scheduling and Integrated Production Scheduling. All these problems can be ranked as combinatorial optimization problems. Combinatorial optimization problems are always NP-hard problems. Its difficult to solve these problems in acceptable time, using traditional optimization methods. But there are still some rapid methods, which are described as intelligent optimization algorithms, help us to get acceptable solution in much more shorter time. Steel-making, Continuous Casting, Hot Rolling are the three important working procedure in Steel Manufacturing Enterprise. Hot Rolling planning is an important activity. Its an combinatorial optimization problem. Hot Rolling plan is not only tightened related to Hot Rolling, but also related to Steel-making, Continuous Casting and many other activities. Schedulers built Hot Rolling Plan by hands formerly. But it always took long time for them to build only one not good plan. Nowadays, advanced steel companies in the world are all using modeling techniques to solve their problem. By this way, they can build more and better plans in much shorter time. The paper studied the Hot Rolling Planning problem of BaoSteel. After thoroughly studied the constraint of BaoSteels Hot Rolling activity, we set up a plan making model, which actually can be classified as a MTSP model. And then we give a intelligent optimization algorithm, which we called Genetic Simulate Annealing Algorithm, to solve the model. 上海交通大学硕士学位论文 2The paper described the Hot Rolling Planning and Optimization Application System of the Automation department of BaoSteel. The application has many positive effects: ? Improve working efficiency, reduce working intensity ? Improve the planning rationality ? Balancing logistics, improve production efficiency ? Improve Hot charge rate, reduce average cost Key Words: Combinatorial Optimization, Modeling & Optimization, Steel-Making, Continuous Casting, Hot Rolling, Hot Rolling Planning, Generic Algorithm, Simulate Annealing Algorithm, Travel Salesman Problem, Application Design.上海交通大学硕士学位论文 1图图 表表 目目 录录 图 1.1 ILOG 优化软件基础结构图.8 图 2.1 钢铁生产工艺流程.16 图 2.2 半连续布置的热连轧机.17 图 2.3 3/4 连续布置的热连轧机.17 图 2.4 全连续布置的热轧机.18 图 2.5 连铸与热轧工序连接方式.19 图 3.1 热轧计划单元组成结构.22 图 3.2 部分映射交叉操作.31 图 3.3 顺序交叉操作.32 图 3.4 循环交叉操作.32 表 3-1 物理退火和组合优化的相似比较.36 图 3.5 遗传算法流程图.36 图 3.6 模拟退火算法流程图.39 图 3.7 遗传模拟退火算法流程图.39 表 3-2 待排板坯数据.41 表 3-3 宽度下降惩罚值设置.43 表 3-4 厚度上升惩罚值设置.43 表 3-5 厚度下降惩罚值设置.