【作者】 方辉；

【导师】 罗阳；

【作者基本信息】 四川大学 ， 机械制造及其自动化， 2003， 硕士

【摘要】 合理利用原材料的优化下料技术，在机械制造、造船、汽车、航空、轻工等行业中的应用非常广泛。国内外研究表明优化下料问题非常困难。该问题作为典型的组合优化难题不但具有NP类复杂度，同时还需要进行复杂图形运算和工艺处理，因此难以建立统一和简单的数学模型，用传统的优化技术来解决。板材排样用于处理平面零件从原材料上的合理下料方案规划，是优化下料问题中应用最广，也是最难的一种。板材排样优化的核心问题是规划零件在板型原材料上最佳的下料组合与每个下料零件在板材上的最优布局方案。探索排样优化技术具有重要的理论意义和工程应用价值。 板材排样问题的核心是解决一个组合优化难题。本文通过分析各种排样优化技术方案的特点，以及在工程应用中排样问题表现出的复杂性、约束性、非线性和建模困难等特点，指出在工程实践中合理的排样优化技术方案，必须适应下料生产作业过程的需要，适合于大规模零件排样，采用具有智能特征的多种算法协同求解的模型来解决下料方案的规划。基于这一技术思想，本文以大型机电产品的毛坯下料生产为背景，研究提出了一种分布式协同优化算法模型，开发了用于大规模矩形零件排样的分布式协同优化排样软件系统。这一排样优化技术方案，能够有效地保证下料生产作业管理中“集中下料”模式的实施，提供最合理的下料方案，同时为计划、调度、材料库存提供决策支持信息。 本文的理论研究与研制开发工作主要包括： (1)首先分析了工业应用领域对排样优化技术的基本要求，并以大型机论文摘要电设备零件下料生产过程为应用背景，建立了排样优化的一般模型。该模型以解决大规模零件优化排样为核心，并可为生产计划、调度、材料库存提供决策支持信息。 (2)在对比分析各种排样优化技术方案的基础上，本文提出了GASA分布式协同优化算法模型。该模型以嵌入模拟退火的遗传算法为基础，利用不同运行参数的多种群并行进化保证种群的基因多样性，并采用结果集交换技术与多种寻优方法协同优化，充分发挥各种方法的局部优势。计算实验的统计结果表明，该模型从整体上提高了求解质量，并在计算效率和求解质量两方面取得了较好的平衡。 (3)本文采用组件技术，提出了在企业内部网络环境中实施大规模矩形零件排样的分布式协同优化计算的技术方案，开发了多个基于DCOM标准的组件，主要包括:优化任务调度组件、排样优化组件集、数据管理组件以及通信管理组件等。 (4)设计开发了大规模矩形零件优化排样系统，主要由四部分组成:协同优化排样、通信管理、结果集管理、优化任务调度。系统界面直观明了，操作简单。更多还原

【Abstract】 Optimizing blank nesting plan, which is mainly aimed on how to make fully use of row material, is widely consist in all kinds of industry such as mechanical machining, shipbuilding, automobile manufacturing, aviation and light industry, etc. Blank nesting plan optimization, also named two dimensional cutting stock problem, is extremely hard because this issue not only possess the complexity of NP hard, but also including intricacy graphic operation and planning disposal. So it’s hard to figure out with conventional optimal approaches based on uniform and simple mathematics model. Blank nesting plan is one of the most hardy and extensively used cutting stock problems, with which manage the two dimension parts’ layout on raw and processed material. The kernel of this problem is to programming the best nesting combination and the best composition scheme of each part on the blank sheet. It has extremely theoretics and engineering application importance to quest for the optimal approaches of blank nesting plan.The nucleus of two dimensional cutting stock problem is to resolve a assembled optimal problem. As the result of the problems’ complexity, restriction, nonlinearity and hard to modeling as well as the analyse of the characteristic of some kinds of optimization model, this paper figured out that the reasonable blank nesting plan must be fit for the demand of roughcast production process and large scale cutting stock based on intelligence multi arithmetic cooperating. Take the producing process of welding row parts of electromechanical equipments as the application background, this thesis brings forward a kind of distributing cooperating optimize arithmetic model based on the technology ideology mentioned above as well as opened up a distributing cooperating cutting stockThe research work supported by the project "Agile workshop manufacturing process intelligence scheduling and its application in electromechanical equipments’ cutting stock process" , which funded by the Youth Science and Technology Research Fund of Sichuan University.system for large scale blank nesting plan which is effectively ensure the operation of convergence cutting stock in blank nesting plan management. The system provided the most reasonable blank nesting plan, as well as the supporting information about planning, scheduling and material storage.The theoretic and development pursuit of this thesis including the following items:1. In the first place, this thesis constitute the currently model of cutting stock problems be based upon the analysis of the demands of industrial application towards the optimization of blank nesting plan combine the producing process of welding row parts of electromechanical equipments as the application background. Regard the two dimensional cutting stock problem as the key problem, this model also providing the decision-making supporting information about production propose, scheduling and raw material storage.2. Brought forward the GASA distributed cooperating model based on the fully and deeply analyse of all kinds of intelligence optimal approaches. Based on genetic algorithm which is embedded simulated annealing, this model ensured the diversity of gene by make use of multi-genus’ parallel evolution with differ functional parameter. The algorithms’ local optimal capability improved by cooperating optimization and acquired the balance of better quality and higher efficiency of large scale two dimensional cutting stock problems.3. This thesis adopt the technic of Component Object Model and reckon that the large scale blank nesting plan system should be operated in the distributed accounting environment of enterprises. The blank nesting plan optimal system mainly including a sorts of components such as optimal assignment scheduling module, optimize module, data management module and the correspondence management module.4. Four functional parts ensured the blank nesting plan system work properly: cooperating optimize approaches, correspondence management, outcome collection and mana更多还原