计算机集成制造系统 ›› 2020, Vol. 26 ›› Issue (第4): 1033-1042.DOI: 10.13196/j.cims.2020.04.017

• 当期目次 • 上一篇    下一篇

基于师生交流机制的改进类电磁机制算法

吴擎1,2,徐惟罡1,张春江3+   

  1. 1.华中农业大学工学院
    2.华中农业大学农业部长江中下游农业装备重点实验室
    3.华中科技大学数字制造装备与技术国家重点实验室
  • 出版日期:2020-04-30 发布日期:2020-04-30
  • 基金资助:
    国家自然科学基金青年基金资助项目(61603145,51905199)。

Teaching-learning based electromagnetism-like mechanism

  • Online:2020-04-30 Published:2020-04-30
  • Supported by:
    Project supported by the National Natural Science Foundation for Youth,China(No.61603145,51905199).

摘要: 为有效解决约束优化问题,提出一种基于师生交流机制的改进类电磁机制算法(TLEM)。算法在继承了类电磁机制(EM)算法良好收敛性的基础上,简化了合力计算过程,引入了移动概率修正移动公式;结合教学优化算法(TLBO),提高了算法的局部搜索精度;改进了EM算法的电量计算公式,并采用一种可行域与支配规则来处理约束。TLEM算法与其他约束优化算法在13个标准测试函数上进行了仿真对比与分析,结果表明TLEM算法的求解能力和稳定性均优于其他算法。将该算法应用于压力容器、压缩弹簧及焊接梁设计问题,验证了TLEM算法在工程设计优化问题上的可行性和优越性。

关键词: 类电磁机制算法, 师生交流机制, 约束优化, 工程设计优化

Abstract: In order to solve constrained optimization problems effectively,a Teaching-Learning based Electromagnetism-like Mechanism(TLEM)algorithm was proposed.On the basis of inheriting the good convergence of the EM algorithm,the force calculation process wes simplified,and the moving probability was introduced to modify the movement formula.By combining with the Teaching-Learning Based Optimization(TLBO)algorithm,the local search accuracy was improved actually.The charge calculation formula of EM algorithm was improved,and a feasibility and dominance rule was adopted to deal with the constraints.The TLEM algorithm was compared and analyzed with other popular constrained optimization algorithms on 13 standard test functions.Experimental results showed that TLEM was superior to other algorithms in solving ability and stability.The proposed algorithm was applied to pressure vessel design,compression spring design and welding beam design,which verified the feasibility and superiority of TLEM algorithm in engineering design optimization.

Key words: electromagnetism-like mechanism algorithm, teaching-learning mechanism, constrained optimization, engineering design optimization

中图分类号: