基于工业背景的矩形板材排样优化研究

发布时间：2018-04-03 02:01

本文选题：矩形板材排样　切入点：遗传算法　出处：《电子科技大学》2015年硕士论文

【摘要】：板材排样优化在现代工业生产中有着广泛的应用,而矩形板材排样优化又是其中最为常见的一类问题。给定一定数量的矩形板材以及目标零件的尺寸与数量,矩形板材排样优化需要找出生产成本最低的切割方案。而生产成本的高低主要由原料使用多少、切割加工难度决定。根据加工工艺的不同,矩形板材排样问题又可以分为有“一刀切”约束和无“一刀切”约束这两类,本文研究的是有“一刀切”约束的矩形板材排样优化问题,其典型的应用场景包括玻璃、石材、木材等板材的生产加工。本文首先介绍了目前排样优化问题的国内外研究现状,指出了针对石材行业多类型、多尺寸原料输入情况下设计排样优化算法的背景与意义,然后分析了在现有加工工艺下矩形板材排样的约束条件,将“易于切割”进行量化并加入到了待优化的目标函数,从而弥补了以往仅仅靠出材率评价排样方案的不足,最终建立了相应的数学模型。接着介绍了矩形板材排样问题已有的经典算法,着重介绍了剩余矩形算法。接下来研究了遗传算法在排样优化问题中的应用,提出了将原料加入染色体编码的方案,从而解决了排样优化中的多规格原料输入问题;使用数学模型中的目标函数合成非线性的适应度评价函数;设计符合排样优化的染色体交叉算子。然后分析了前面设计的遗传算法在同一规格的原料、零件数量较多时存在的不足,提出了同类复用策略,并尝试将其融入到遗传算法中,最终使用贪心算法实现同类复用策略,并通过分阶融合的方式将其与遗传算法结合使用,从而解决了遗传算法排样结果与实际生产习惯不够契合的问题。在此基础上,提出了进一步改进排样算法的整体摆放策略,并将其应用到贪心算法当中。对于未充分利用的原料,提出了二分搜索策略来提高余料的可用性。接下来使用前面提出的排样算法设计并编码实现了针对石材行业的自动排样系统,将其分为UI模块和算法模块两部分,使其具有较高的可维护性,同时加入了一些石材行业的常用功能,提高了排样系统的实用性。最终,我们通过一个符合实际生产需要的具有多类型、多规格原料与零件的测试用例对自动排样系统进行测试,成功计算出所有220个目标零件的排样方案,整体出材率为94%,同时排样结果也具有同类复用、整体摆放的特征,达到了排样算法的设计目标。
[Abstract]:Plate layout optimization is widely used in modern industrial production, and rectangular plate layout optimization is one of the most common problems.Given the size and quantity of a certain number of rectangular plates and target parts, the cutting scheme with the lowest production cost is needed to optimize the layout of rectangular plates.The cost of production is mainly determined by the amount of raw materials used and the difficulty of cutting and processing.According to the different processing technology, the rectangular plate layout problem can be divided into two types: "one size fits all" constraint and no "one size cut" constraint. In this paper, the rectangular plate layout optimization problem with "one size fits all" constraint is studied.Typical application scenarios include the production and processing of glass, stone, wood, etc.This paper first introduces the current research status of layout optimization problem at home and abroad, and points out the background and significance of designing layout optimization algorithm under the condition of multi-type and multi-size raw material input in stone industry.Then, the constraint conditions of rectangular plate layout under the existing processing technology are analyzed, and the "easy to cut" is quantified and added to the objective function to be optimized.Finally, the corresponding mathematical model is established.Then the classical algorithms of rectangular plate layout are introduced, and the residual rectangle algorithm is emphasized.Then, the application of genetic algorithm in layout optimization is studied, and the scheme of adding raw material to chromosome coding is put forward to solve the problem of multi-specification raw material input in layout optimization.The objective function in the mathematical model is used to synthesize the nonlinear fitness evaluation function, and the chromosome crossover operator is designed according to the layout optimization.Then it analyzes the shortcomings of the genetic algorithm in the same specification and the large number of parts, and puts forward the similar reuse strategy, and attempts to integrate it into the genetic algorithm, and finally uses greedy algorithm to realize the same reuse strategy.The genetic algorithm is combined with genetic algorithm in order to solve the problem that the layout result of genetic algorithm is not consistent with the actual production habit.On this basis, the overall placement strategy of the improved layout algorithm is proposed and applied to the greedy algorithm.A binary search strategy is proposed to improve the availability of raw materials.Then, the automatic layout system for stone industry is designed and coded with the proposed layout algorithm, which is divided into two parts: UI module and algorithm module, so that it has high maintainability.At the same time, some common functions of stone industry are added to improve the practicability of the layout system.Finally, we test the automatic layout system with a test case of multi-type, multi-specification raw materials and parts that meet the actual production needs, and successfully calculate the layout scheme of all 220 target parts.The overall output rate is 94 and the layout result has the characteristics of similar reuse and overall placement, which achieves the design goal of the layout algorithm.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TB30;TP18

【参考文献】