CIESC Journal ›› 2018, Vol. 69 ›› Issue (7): 3135-3148.DOI: 10.11949/j.issn.0438-1157.20171625

Previous Articles     Next Articles

A structure evolution strategy motivated by large step size for optimization of heat exchanger network

SUN Tao, CUI Guomin, CHEN Jiaxing   

  1. Institute of New Energy Science and Technology, University of Shanghai for Science and Technology, Shanghai 200093, China
  • Received:2017-12-11 Revised:2018-03-21 Online:2018-07-05 Published:2018-07-05
  • Supported by:

    supported by the Capacity Building Plan for Some Non-military Universities and Colleges of Shanghai Scientific Committee (16060502600) and the National Natural Science Foundation of China (51176125).

一种大步长激励的结构进化策略应用于换热网络优化

孙涛, 崔国民, 陈家星   

  1. 上海理工大学新能源科学与工程研究所, 上海 200093
  • 通讯作者: 崔国民
  • 基金资助:

    上海市科委部分地方院校能力建设计划项目(16060502600);国家自然科学基金项目(51176125)。

Abstract:

Random walk algorithm with compulsive evolution (RWCE) was an effective method for the optimization of heat exchanger network with the advantage of simple evolutionary strategy and less control parameters. The RWCE algorithm had the capacity to jump out of the local optimal through continuously accepting the imperfect solutions, but its jumping efficiency was often limited by the step size. The global search ability of the algorithm was reduced. Therefore, a novel combined step size generation method was established by studying the step size distribution of RWCE algorithm and analyzing the mechanism of large step size assisted to jump out of local optimal, then a structure evolution strategy motivated by large step size was proposed. Finally, the validity of the proposed structure evolution strategy was verified by three cases. Compared with the literature results, the results showed that the structure evolution strategy motivated by large step size had a strong global search ability.

Key words: process systems, algorithm, global optimization, step size distribution, structure evolution

摘要:

强制进化随机游走算法(random walk algorithm with compulsive evolution,RWCE)具有进化策略简单、控制参数少、全局搜索能力强等特点,是换热网络优化的一种有效方法。鉴于步长对RWCE算法跳出局部最优解有重要的影响,针对RWCE算法的步长分布进行研究,通过分析大步长在局部极值跳出中的作用机理,建立了一种全新的步长组合生成方法,并据此提出了一种大步长激励的结构进化策略。最后,通过3个算例验证了结构进化策略的有效性,与文献结果进行对比,结果表明大步长激励的结构进化策略提高了算法的全局搜索能力。

关键词: 过程系统, 算法, 整体优化, 步长分布, 结构进化

CLC Number: