基于遗传算法的水火弯板工艺参数预报方法研究  

Investigations of technique parameter prediction method of line heating based on genetic algorithm

作　　者：刘玉君[1] 郭培军[1] 纪卓尚[1] 邓燕萍[1] 

机构地区：[1]大连理工大学船舶工程学院,辽宁大连116024

出　　处：《大连理工大学学报》2006年第2期235-240,共6页Journal of Dalian University of Technology

基　　金：国家"八六三"高技术研究发展计划资助项目(2001AA421200)

摘　　要：以钢板加工时间和钢板收缩率曲线拟合优度作为两个目标函数,建立水火弯板工艺参数优化问题数学模型,并应用权重系数变化法,将问题转化为单目标优化问题.基于分层遗传算法(HGA),应用三次样条插值、误差分析等方法,将目标函数转化为群体中个体的适应度函数.由加热线条数、相邻加热线之间距离(加热线间距)和加热线收缩量作为3个约束条件,构造分层遗传算法编码,并且内外层编码在遗传运算时分别采用不同的遗传算子,最终得到钢板水火弯板加工工艺参数预报结果.经算例验证,该方法可以对船体外板作出良好的水火弯板工艺参数预报.Plate processing time and goodness of fit of shrinkage ratio curve are taken as two objective functions to build the mathematical model of the optimization issue of line heating technology parameters, and then the two objective optimization problems are translated into one by changing power coefficients. Based on hiberarchy genetic algorithm （HGA）, the interpolation function with cubic spline, error analysis, etc. is used to change the objective function into individual fitness function of colony. Combining three constraints, which include the number of heating lines, the space between lines and the shrink quantity of line, HGA coding is constructed, and the inner coding and outer coding adopt different heredity arithmetic operators in inherit operating. Finally, the optimization issue of line heating technology parameters is solved. It is proved that the method can give fine prediction of line heating technique parameter for steel shell plate.

关 键 词：水火弯板 工艺参数预报 分层遗传算法 目标函数 

分 类 号：U671.3[交通运输工程—船舶及航道工程]