遗传算法优化乘积混料试验最佳配比条件的研究  

A Study of Genetic Algorithm to Optimize the Best Proportion and Processing Conditions of Product Mixture Design

作　　者：赵磊[1] 吴小娟[1,2] 任雯[1] 王祥[1] 王华芳[1] 魏珍[1] 仇丽霞[1] 

机构地区：[1]山西医科大学公共卫生学院卫生统计学教研室,030001 [2]滨州医学院附属医院

出　　处：《中国卫生统计》2016年第2期205-208,共4页Chinese Journal of Health Statistics

基　　金：山西省自然科学基金项目(2013011059-3)

摘　　要：目的评价遗传算法在乘积混料试验设计的实际研究中的优化效果。方法利用Expert Design软件对混料设计的数据建立回归模型,并利用Genetic Algorithm Platforms v1.0软件进行单目标优化,并利用Mintab16.0软件绘制等高线图。将遗传算法所得结果与传统的等高线法结果进行比较。结果通过遗传算法求得最佳配比分别为:氧化铷为52.4%,氧化锡为47.6%,氧化钛为0%,且煅烧条件在500摄氏度时,焊条的使用寿命达到32.60h。使用寿命比等高线法得到的最优值略高,说明遗传算法能够找到试验中的最佳配比及工艺条件。结论对于乘积混料试验设计利用遗传算法进行单目标优化效果理想,可进一步应用于医药学领域。Objective To evaluate the optimizations of product mixture design based on genetic algorithm. Method Using expert design software to establish a regression model,using Genetic Algorithm Platforms v1. 0 software to achieve single objective optimization with the data from mixture design and using Mintab16. 0 software to draw contour plot. Compare the result from genetic algorithm and contour method. Results The best proportion calculated by genetic algorithm are rubidium oxide52. 4%, tin oxide 47. 6%, titanium oxide 0% and the condition of calcinations at 500℃. The electrode service life reach up to32. 60 h, the service life is a little higher that contour method. It indicates that using genetic algorithm can find the optimum proportion and technological conditions in product mixture design. Conclusion For product mixture design,using genetic algorithm to achieve single objective optimization, the effect is satisfactory. The method can be extended applied into the medicine field.

关 键 词：乘积混料试验 遗传算法 单目标优化 最佳配比 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] R91[自动化与计算机技术—控制科学与工程]