解决矩形件带排样问题一种遗传算法.docxVIP

广西师范大学硕士研究生学位论文 种采用两个种群并行进化的遗传算法求解 RSPP 的新算法。矩形件排样问题中如何产生 排样序列是一个关键的问题，许多学者应用遗传算法，模拟退火算法等算法产生排样序 列，再将产生的排样序列与某种矩形件近似排样算法结合求解矩形件排样问题，取得了 较好的排样效果。多种群遗传算法是在基本遗传算法基础上改进的一种算法。本文分析 了多种群遗传算法，应用两个规模相当的子种群进行并行进化，其中两个子种群采用不 同的初始策略。产生两种不同性质的矩形零件递减序列，并设计了求解 RSPP 的相关操 作，定义了求解 RSPP 的适应度函数。将遗传算法产生矩形件的递减序列，按照本文的 基于最低水平线的择优插入算法和本文的适应度函数得到序列的适应度值，并生成排样 图。通过矩形零件序列的适应度值的比较，从而得到一个较优的排样方案。 然后，规划和设计了排样系统的基本功能模块，开发了一个基于遗传算法的矩形件 优化排样系统。通过大量实验测试，并将实验结果与应用遗传算法、模拟退火算法、随 机算法求解矩形件排样问题的实验结果进行了比较和分析，验证了该系统的算法的有效 性。 最后，论文对己完成的工作进行了总结，提出了需要进一步改进的工作。 关键词：矩形毛坯排样，遗传算法，两维切割，切割下料，优化 II 广西师范大学硕士研究生学位论文 A genetic algorithm for the rectangular strip packing problem  Student: Zhao Xinfang  Tutor: Cui Yaodong  Major: Computer Application and Technique Research Area: Optimization and Computation Technology; Computer added design Grade: 2005 Abstract The rectangle packing problem exists widely in many industries, such as the mechanical, furniture and costume industries. Solving this problem can save raw materials, streamline production processes, reduce production cost and increase the efficiency of industry. Irregular layouts can be transformed into rectangular packing problems by computer graphics processing technology. The rectangle strip packing problem (RSPP) can be stated as putting some quantified rectangles P , P ,…, P on the given sheet, which is fixed with width and isn’t fixed with height, with an aim to minimizing sheet’s unused area. RSPP is an importance branch of computer aid nesting. But it is a high computational complexity NP-complete problem in theory. So when the scale of the problem is large, we can not use exact algorithm to acquire the optimal solution. Therefore, the study has important practical and theoretical value. The genetic algorithm is global search optimization compute technology based on biology anagenesis and random selection. It simulates biology anagenesis basal progress, uses the Digital gene string to analog the chromosome, and imitate the basal biology evolution pro