二维矩形件排样问题的自适应多岛遗传算法优化  

Adaptive multi-island genetic algorithm optimization on layout problem for 2D rectangular parts

作　　者：曾晓亮[1] 吴琼[1] 袁旭华[2] Zeng Xiaoliang;Wu Qiong;Yuan Xuhua(College of Design Engineering,Jiangxi College of Applied Technology,Ganzhou 341000,China;College of Mathematics and Computer Science,Yan′an University,Yan'an 716000,China)

机构地区：[1]江西应用技术职业学院,设计工程学院,江西赣州341000 [2]延安大学数学与计算机科学学院,陕西延安716000

出　　处：《锻压技术》2020年第12期53-58,共6页Forging & Stamping Technology

基　　金：江西省高等学校教学改革研究重点课题(JXJG-16-52-2)。

摘　　要：为了提高二维矩形件排样问题的板材利用率、降低生产成本,提出了基于自适应多岛遗传算法的排样优化方法。使用六元数组对矩形件进行表征,以板材利用率最大为优化目标,建立了排样问题的带约束优化模型。以多岛遗传算法为基础,引入了交叉概率和变异概率的自适应调整方法,从而提出了自适应多岛遗传算法;针对排样问题的特殊性,对遗传算子进行适应性设计,提出了环形交叉方法和交换变异策略,保证了执行遗传算子前、后的矩形规模不变;提出了最低水平线启发式算法的基因解码方法。使用规模为30和59的两组矩形件排样实验进行验证,结果表明:与分布估计排样方法相比,自适应多岛遗传算法的排样结果的板材利用率更高,且排样方法的稳定性也优于分布估计排样方法。The optimization method of layout based on adaptive multi-island genetic algorithm was proposed in order to increase the utilization rate of plate and reduce production cost for the layout problem of 2 D rectangular parts.Then,the rectangular part was characterized by six-element array,taking the maximum utilization rate of plate as the optimizing goal,the optimization model with constraint of layout problem was built.Based on multi-island genetic algorithm,the adaptively adjusting method of crossover probability and mutation probability was introduced,and the adaptive multi-island genetic algorithm was put forward.Furthermore,for the particularity of layout problem,the genetic operators were designed adaptively,and the circular crossover method and changing mutation strategy were given to ensure that the rectangle scale remained unchanged before and after the execution of genetic operator.Finally,the gene decoding method was set by the lowest horizontal line heuristic algorithm,and two sets of layout experiments for rectangular part with the scales of 30 and 59 for verification were conducted.The result shows that compared with the distribution estimation layout method,the utilization rate of plate and the stability of layout method for the layout results of adaptive multi-island genetic algorithm are higher.

关 键 词：矩形件排样 自适应多岛遗传算法 启发式算法 基因解码 板材利用率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]