引用本文:李静,王京,杨磊,刘森.热处理炉钢板温度的自适应混沌粒子群算法–最小二乘支持向量机优化预报算法[J].控制理论与应用,2011,28(12):1825~1830.[点击复制]
LI Jing,WANG Jing,YANG Lei,LIU Sen.Optimized prediction algorithm with adaptive chaos particle swarm optimization–least squares support vector machine for steel plate temperature prediction in heat treatment furnace[J].Control Theory and Technology,2011,28(12):1825~1830.[点击复制]
热处理炉钢板温度的自适应混沌粒子群算法–最小二乘支持向量机优化预报算法
Optimized prediction algorithm with adaptive chaos particle swarm optimization–least squares support vector machine for steel plate temperature prediction in heat treatment furnace
摘要点击 2742  全文点击 1312  投稿时间：2010-08-03  修订日期：2011-02-15
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/j.issn.1000-8152.2011.12.CCTA100895
  2011,28(12):1825-1830
中文关键词  热处理炉  粒子群优化算法  支持向量机  混沌
英文关键词  heat treating furnace  PSO(particle swarm optimizer algorithm)  SVM(support vector machine)  chaos
基金项目  “十一五”国家科技支撑计划资助项目(2006BAE03A06).
作者单位E-mail
李静* 北京科技大学 冶金工程研究院 lijing6332@ustb.edu.cn 
王京 北京科技大学 冶金工程研究院  
杨磊 舞阳新宽厚钢板责任有限公司轧钢厂  
刘森 北京亿玮坤节能科技有限公司,  
中文摘要
      针对传统传热模型参数调整较复杂和模型精度较低的问题, 构建了一种基于改进粒子群算法优化最小二乘支持向量机(least squares SVM, LSSVM)的钢板温度预报模型. 首先, 对基本粒子群算法进行分析，提出自适应混沌粒子群算法(adaptive chaos PSO, ACPSO), 并通过性能指标定量评价验证算法的有效性、鲁棒性和寻优效率. 其次, 采用LSSVM建立钢板温度预报模型, 并选用径向基函数作为核函数, 用ACPSO算法优化该模型参数. 最后, 结合现场数据进行仿真研究和工程应用, 结果表明基于该算法建立的钢板温度预报模型具有较高的预报精度, 达到智能调优的目的.
英文摘要
      To deal with the difficulty in parameter adjustment and the low precision of the traditional heat-conduction model, we build a prediction model for the steel plate temperature, based on the least-squares-support-vector machine(LSSVM) which is optimized by the improved particle-swarm algorithm. First, on the basis of the particle-swarm algorithm, we propose an adaptive chaotic particle-swarm algorithm(ACPSO) for which the validity, robustness and the optimization efficiency are quantitatively evaluated based on performance indices; and then, the radial basis functions are selected as the kernel function. Thus, the temperature prediction model of steel plate is built with LSSVM and optimized with ACPSO algorithm. Finally, the model is simulated by using the data acquired from the site and used in practical operation; the result indicates that the prediction model based on ACPSO and LSSVM has higher prediction accuracy than the tradition one, achieving the goal of intelligent optimization.