计算机应用 ›› 2015, Vol. 35 ›› Issue (11): 3126-3129.DOI: 10.11772/j.issn.1001-9081.2015.11.3126

• 2015年全国开放式分布与并行计算学术年会(DPCS 2015)论文 • 上一篇    下一篇

基于K-均值的“教”与“学”优化算法

黄祥东1, 夏士雄1, 牛强1, 赵志军2   

  1. 1. 中国矿业大学 计算机科学与技术学院, 江苏 徐州 221116;
    2. 舟山市定海区交通建设事务中心, 浙江 舟山 316000
  • 收稿日期:2015-06-17 修回日期:2015-07-06 发布日期:2015-11-13
  • 通讯作者: 夏士雄(1961-),男,辽宁抚顺人,教授,博士,主要研究方向:智能控制、数据挖掘、工业通信网络.
  • 作者简介:黄祥东(1990-),男,山东邹城人,硕士研究生,主要研究方向:人工智能、机器学习; 牛强(1974-),男,河南南阳人,教授,博士,主要研究方向:数据挖掘、智能优化算法、智能信息处理; 赵志军(1966-),男,浙江余姚人,政工师,主要研究方向:人工智能、传感器网络.
  • 基金资助:
    江苏省产学研联合创新资金前瞻性联合研究项目(BY2014028-09);国家海洋局数字海洋科学技术重点实验室开放基金资助项目(KLDO201304);浙江省交通运输厅科研计划项目(2014T25).

Improved teaching-learning-based optimization algorithm based on K-means

HUANG Xiangdong1, XIA Shixiong1, NIU Qiang1, ZHAO Zhijun2   

  1. 1. School of Computer Science and Technology, China University of Mining and Technology, Xuzhou Jiangsu 221116, China;
    2. Ministry of Transport of Dinghai District, Zhoushan Zhejiang 316000, China
  • Received:2015-06-17 Revised:2015-07-06 Published:2015-11-13

摘要: 在解决复杂多峰优化问题时,传统的"教"与"学"优化算法易于陷入局部搜索且优化效率较低.针对此问题,提出了一种基于K-均值的"教"与"学"优化改进算法,算法采用K-均值来降低种群规模,又针对"教"和"学"两个阶段进行相应改进,提高全局收敛速度;还加入了"变异"操作来避免算法陷入局部最优.实验对7个单峰值优化问题和2个有代表性的多峰值优化问题进行优化,并与手榴弹爆破算法和传统"教"与"学"优化算法进行比较,实验结果表明,该改进算法在单峰和多峰测试函数中,均能快速高效地寻得全局最优解,优于原始"教"与"学"优化算法.

关键词: "教"与"学"优化算法, K-均值, 多峰函数, 全局最优解

Abstract: For the complex multimodal optimization problems, the traditional Teaching-Learning-Based Optimization (TLBO) algorithm is easy to fall into local search and has low optima efficiency. In order to solve the problem, an improved TLBO algorithm based on K-means was proposed in this paper. It used the K-means to decide the population into small populations for reducing the population size and correspondingly improved the "teaching" and "learning" stages to improve the speed of global convergence. At the same time, the proposed algorithm added "mutation" operation to avoid the local optimum. In the experiments, seven unimodal and two multimodal optimization problems were optimized by the algorithm proposed in this paper. The optimization results were compared grenade explosion method and traditional TLBO algorithm. The experimental results show that the improved algorithm can quickly and efficiently find the globally optimal solution in both unimodal and multimodal functions and the improved algorithm is better than the traditional TLBO algorithm in the ability of searching the globally optimal solutions.

Key words: Teaching-Learning-Based Optimization (TLBO) algorithm, K-means, multimodal functions, global optimal solution

中图分类号: