基于迁移学习灰支持向量回归机的交互式进化计算
作者:
作者单位:

郑州航空工业管理学院

作者简介:

通讯作者:

中图分类号:

TP301

基金项目:


Interactive Evolutionary Computation Based onTransfer Learning Grey Support Vector Regression
Author:
Affiliation:

Zhengzhou University of Aeronautics

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    机器感知评价和种群进化策略是交互式进化计算的2个核心问题。针对这2个问题, 提出基于迁移学习灰支持向量回归机的个体适应值预测方法和聚类进化策略。迁移学习灰支持向量回归机对根据个体相似性估计的适应值预测,提高评价精度。提出基于个体平均距离的聚类选择算子,克服NSGA-II只选择拥挤度大的个体带来的局部最优收敛问题。提出基于个体平均距离的交叉算子,扩大子代搜索区域,增强种群多样性。基于上述策略,采用NSGA-II范式实现交互式进化计算。将该算法应用于室内灯光调色问题,验证所提方法的有效性。

    Abstract:

    Machine perception evaluation and population evolution strategy are core issues for interactive evolutionary computation(IEC).This study put forward a fitness prediction method based on transfer learning grey support vector regression(TG-SVR) and clustering evolution strategy for these issues. TG-SVR can forecast fitness estimated according to individual similarity to improve evaluation precision.A clustering selection operator is proposed based on individual average distance,which can overcome the local convergence problem because NSGA-II only choose individuals with large crowding distance.A clustering crossover operator is put forward based on individual average distance,which can expand the offspring search area and enhance population diversity.Based on above strategies, an interactive evolutionary computation is employed to NSGA-II.The proposed interactive evolutionary computation base on transfer learning grey support vector regression(IEC-TGSVR)is applied to interior light toning optimization problem, and its outstanding performance is experimentally demonstrated.

    参考文献
    相似文献
    引证文献
引用本文
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2020-04-14
  • 最后修改日期:2020-07-29
  • 录用日期:2020-08-04
  • 在线发布日期:
  • 出版日期: