计算机科学与探索 ›› 2022, Vol. 16 ›› Issue (6): 1362-1373.DOI: 10.3778/j.issn.1673-9418.2010049

• 人工智能 • 上一篇    下一篇

融入社会心理学理论的教与学优化算法

何佩苑(), 刘勇   

  1. 上海理工大学 管理学院,上海 200093
  • 收稿日期:2020-10-19 修回日期:2021-01-07 出版日期:2022-06-01 发布日期:2021-01-25
  • 通讯作者: + E-mail: 17805058112@163.com
  • 作者简介:何佩苑(1997—),女,江苏溧阳人,硕士研究生,主要研究方向为智能优化、系统工程。
    刘勇(1982—),男,江苏金湖人,博士,副教授,硕士生导师,主要研究方向为智能优化、服务网络设计与优化、系统工程。
  • 基金资助:
    上海市“科技创新行动计划”软科学研究重点项目(18692110500);上海市社会科学规划课题(2019BGL014);上海市高原科学建设项目(第2期);上海理工大学科技发展资助项目(2020KJFZ040)

Teaching-Learning-Based Optimization Algorithm with Social Psychology Theory

HE Peiyuan(), LIU Yong   

  1. School of Management, University of Shanghai for Science and Technology, Shanghai 200093, China
  • Received:2020-10-19 Revised:2021-01-07 Online:2022-06-01 Published:2021-01-25
  • About author:HE Peiyuan, born in 1997, M.S. candidate. Her research interests include intelligent optimization and system engineering.
    LIU Yong, born in 1982, Ph.D., associate profes-sor, M.S. supervisor. His research interests include intelligent optimization, design and optimization of service network and system engineering.
  • Supported by:
    Key Project of Soft Science Research in Shanghai Science and Technology Innovation Action Plan(18692110500);Social Science Planning Project of Shanghai(2019BGL014);Plateau Science Construction Project of Shanghai (NO.2);Science and Technology Development Funding Project of University of Shanghai for Science and Technology(2020KJFZ040)

摘要:

教与学优化算法(TLBO)是一种模拟教学过程的启发式优化算法。针对TLBO算法寻优精度低、稳定性差的特点,提出了基于社会心理学理论改进的教与学优化算法(SPTLBO)。该算法在改进中考虑了人的心理因素:在原TLBO算法的“教”阶段中结合社会心理学的“期望效应”理论,教师对高期望学生采取一对一教学策略,使得优秀学生更快向教师靠近;为了保留学生的多样性,学生依据认知风格可分为“场独立”与“场依存”两种类型,不同类型的学生将采取不同的交流方式进行学习;在“教”“学”阶段后,结合自我调节理论,学生进入学习方法调整阶段,从而增强了自我探索能力,提高学生整体水平。此外,引入自适应学生更新因子,模拟环境对学生学习效率的影响,增加算法的全局搜索能力,避免出现在初期迭代中陷入局部最优的情况。在25个标准测试函数上进行实验,结果表明SPTLBO算法相较基本TLBO算法和其他智能优化算法,在寻优精度和收敛速度方面都更具优势。

关键词: 教与学优化算法(TLBO), 社会心理学, 期望效应, 场独立-场依存, 自我调节理论, 自适应更新因子

Abstract:

Teaching-learning-based optimization (TLBO) is a heuristic optimization algorithm that simulates the teaching process. In view of the low precision and poor stability of TLBO algorithm, an improved teaching-learning-based optimization algorithm named SPTLBO (social psychology teaching-learning-based optimization) is proposed. Human psychological factors are considered in the improvement of the algorithm. In the “teaching” stage, combining the “expectation effect” theory in social psychology, teachers adopt one-to-one teaching strategy for students with high expectations, which makes outstanding students approach teachers faster. According to cognitive style, students can be divided into two types, “field independence” and “field dependence”, so that it can preserve the diversity of students. Different types of students will adopt different communication methods to learn. After the “teaching” and “learning” stages, combined with the theory of self-regulation, students enter the stage of learning method adjustment. It can enhance the ability of self-exploration and improve the overall level of students. In addition, an adaptive student update factor is introduced to simulate the influence of environment on students’ learning efficiency, which increases the global search ability of the algorithm and avoids falling into local optimum in the initial iteration. The test of 25 test functions shows that, compared with the basic TLBO algorithm and other intelligent optimization algorithms, the SPTLBO algorithm has more advantages in the optimization accuracy and convergence speed.

Key words: teaching-learning-based optimization algorithm (TLBO), social psychology, expectation effect, field independence-field dependence, self-regulation theory, adaptive update factor

中图分类号: