摘要
为了提高零件平面度测量结果,采用改进多元宇宙算法。首先建立零件平面度测量模型,设置非线性最优化问题的目标函数;接着基于对数函数对多元宇宙算法虫洞概率调整,调整后的虫洞概率随着迭代次数的增加,物体穿越虫洞的机会增加,从而加速了宇宙物体的移动,提高了算法的搜索性能;然后基于指数函数对宇宙算法移动步长优化,在迭代初期具有较大值使得宇宙能够以较快速度进行全局探索,迭代后期具有较小值使得宇宙能够慢速进行局部开发,提升算法探索、开发能力;最后给出了算法流程。实例验证结果表明改进多元宇宙算法测量误差平均值为5.212μm,相比LSM、PSO、GA、DE、ANN算法分别减少了37.63%、33.61%、27.51%、19.41%、10.17%,测量所需时间为3.282 s,分别减少了36.95%、30.33%、26.11%、18.52%、12.74%,明显优于其它算法,测量效果较好。
In order to improve the flatness measurement results of parts,the improved multi universe algorithm is proposed.Firstly,the flatness measurement model of parts is established,and the objective function of nonlinear optimization problem is set.Secondly,the wormhole probability of multi verse optimization is adjusted based on logarithmic function,with the iteration times,the probability of objects crossing wormhole is increased,so that accelerating the movement of verse and improving the search performance.Thirdly,the moving step length of multi verse is optimized based on exponential function,the verse is explored globally fast at the beginning of the iteration using large value,while developed locally slowly using smaller value at the late stage,so that enhancing the ability of exploration and development.Finally,the process is given.The experimental results show that the average measurement error of the improved multi universe algorithm is 5.212μm,which is 37.63%,33.61%,27.51%,19.41%and 10.17%less than LSM,PSO,GA,DE and ANN algorithm respectively.The measurement time is 3.282 s,which is 36.95%,30.33%,26.11%,18.52%and 12.74%less than LSM,PSO,GA,DE and ANN algorithm respectively,it is better than other algorithms obviously.
作者
熊保玉
XIONG Bao-yu(Chengdu Vocational&Technical College of Industry Equipment Manufacturing Institute,Chengdu 610218,China)
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2021年第3期251-256,共6页
Journal of Optoelectronics·Laser
基金
四川省教育厅科研项目(18ZB0046)资助项目。
关键词
测量
平面度
多元宇宙
虫洞概率
移动步长
measurement
flatness
multi verse
wormhole probability
moving step length
