上海交通大学
博士学位论文
皮革制造业中排样关键技术研究
姓名：***
申请学位级别：博士
专业：机械设计及理论
指导教师：***
20040401
上海交通大学博士学位论文 摘要 - I -
皮革制造业中排样关键技术研究
摘 要中班观察笔记
随着我国皮革产业的发展大小
瑕疵孔等质量约束考虑以及多个皮料中优化排样的实现都是非常有挑战性的针对皮料的轮廓形状不仅有直线和圆弧
由此创新地提出了几何信息离散化的处理方式
离散处理中皮料以及样片的几何信息被转化成最小虚拟网格矩形包络的一组网格值信息
将各个小区域定义为0或非
硬笔书法Simulated Annealing Algorithm
傲然挺立的意思在二维不规则形体排样问题上的应用
BA Prompt Simulated Re-annealing Nesting
Algorithm
ΪÁËÑéÖ¤PSRNA 的效率
它不仅实现了较高的排样材料利用率所有的随机扰动能够同时发生以使独立优化参数的数量不至于成为工作的瓶颈
上海交通大学博士学位论文 摘要 - II - 皮料的排样处理时
样片数成为提高排样效率的瓶颈采用不同方式的遗传算法(Genetic Algorith
ÎÄÖÐΪ²»¹æÔòÁ½Î¬Ñù¼þÔÚ²»¹æÔòƤÁÏÉϵÄÅÅÑùÎÊÌ⿪·¢³öλÕóÁÐÅÅÑùÒÅ´«Ëã·¨根据皮革应用特点定义了灵活的遗传算子
它们使搜索质量显著提高
休闲搭配>梭梭柴
其结果大大缩减了优化参数后续搜索中的状态空间
在采用启发式底左算法
提出了一种简洁处理算子方式Simulated Annealing Bad Genetic Nesting Algorithm ÔÚ´óСºñ¶È²»µÈ
ÕâÖÖ×éºÏ·½·¨±È´¿´âµÄSA 或GA 方法具有更突出和优异的性能
使得优化不但效率高而且较快的逼近近似最优解一棵小树作文
还特别适用于二维不规则形体在多个二维不规则平面上的优化排样
²¢ÇÒ¸ù¾ÝÑо¿Êµ¼Ê
上海交通大学博士学位论文 摘要 - III -
得到快速模拟重复退火排样算法
实验
证明其效果优于传统的BA 和CA 方法３但在执行时间上有差距４
很好地处理了变异操作
样片的最优布置顺序
和角度依靠SABGNA 算法来实现由一个满足特定冷却曲线的温度来控制
直到最后搜索收敛为止使得排样能在多个皮料中优选展开
关键词
上海交通大学博士学位论文 ABSTRACT
KEY TECHNIQUE STUDY
古泰拳
带燕的诗句ON NESTING PROBLEMS IN THE LEATHER INDUSTRY
ABSTRACT
The geometrical profiles of leather sheets and stencils are directly obtained using a digitizer or indire
ctly t up using a scanner followed by vector processing. Besides straight lines and arcs, the leather profiles abound in arbitrary curves. According to this feature, we studied the different applications of polygon approximation and further developed a processing scheme of discretizing geometrical information. Thus the whole processing is independent of the geometrical profiles of leather sheets and stencils, which fundamentally solves the problems caud by the irregularity of leather. During the discretization processing, geometrical information is converted to a t of grid values within a minimum virtual grid rectangle. Bad on the pixel theory, small grid gments are defined as 0 or non-0 by investigating the leather profile property within each virtual grid gment. Certain data structures are ud to implement the above information description.
The application of Simulated Annealing (SA) in the nesting problems of two-dimensional geometry is thoroughly studied in the thesis. Bad on the study of Boltzmann Annealing (BA) and Cauchy Annealing (CA), we obtained a Prompt Simulated Re-annealing Nesting Algorithm (PSRNA). According to our experiments, PSRNA is more suitable for the nesting problem of two-dimensional irregular geometry than BA and CA. To test the performance of PSRNA, we implemented and compared the different annealing strategies. The experimental results show that PSRNA is superior to CA and BA in terms of both the cost and the execution time. PSRNA can achieve relatively higher
material usage within the least execution time. In PSRNA, all the random disturbances occur simultaneously, which is beneficial to the arch for an optimal configuration. It is also found that CA performs better than BA. This can partially explain why CA is widely ud in the nesting problems of different fields as a conventional SA approach. However, to avoid the number of independent optimization parameters becoming a bottleneck for an efficient arch, the number of the stencils ud in PSRNA had better be less than 30. In addition, the problem becomes more apparent if this approach is applied in the nesting problem of multiple sheets. Therefore, to improve the nesting efficiency, we continued to study other algorithms.
Then our rearch work on the Genetic Algorithm (GA) realizing the nesting problem of two-dimensional irregular geometry is introduced in this thesis. Bad on the conventional bit-string chromosomal reprentation, we developed a Bit-Array Nesting Genetic Algorithm (BANGA).
- IV -
