摘要
针对复杂换热网络混合整数非线性问题,提出了一种由混沌蚁群算法、局部搜索策略和结构进化策略组成的混合算法,同步综合换热网络。首先采用混沌蚁群算法初步优化换热网络,蚂蚁个体根据混沌搜索机制遍历整个求解域。随后引入Powell法作为局部搜索策略,加强蚂蚁个体的局部搜索能力。最后结合结构进化策略,限制算法的搜索空间,优化蚂蚁个体表示的换热网络结构,并将优化后的信息反馈。蚂蚁会根据自身、邻居和反馈的信息作进一步搜索,直到算法收敛于全局最优解。通过算例对算法进行验证,结果表明,混沌搜索机制使混合算法具有很好的全局搜索能力;Powell法加强了算法的局部搜索能力,提高了求解精度;结构进化策略能够有效地缩减搜索区间,提高搜索效率。所以混合算法能够很好地兼顾处理连续变量和整型变量,适用于换热网络综合。
Simultaneous synthesis of heat exchanger network is primarily regarded as complex mixed integer nonlinear programming models （MINLP）. Therefore, a novel hybrid algorithm which consists of a combination of chaotic ant swarm algorithm （CAS）, structure evolution strategy （SE） and local search strategy was proposed. Firstly, CAS algorithm （which ants can traverse the whole solution space based on chaotic search mechanism） was applied to obtain an initial network configuration. Subsequently, the Powell method was introduced as local search strategy to enhance the solution accuracy and stability. Finally, the structure evolution strategy built the structure represented by the ants and limited the search space. The evolved structures were sent to the CAS algorithm for guiding the next search of ants. Two cases published in the previous literature were examined to evaluate the performance of the presented algorithm. The results show that the obtained values are better than that reported previously. Thus the hybrid algorithm can simultaneously handle both continuous and discrete variables, and the local optimizing ability and the structural search ability are clearly enhanced.
出处
《高校化学工程学报》
EI
CAS
CSCD
北大核心
2016年第6期1380-1390,共11页
Journal of Chemical Engineering of Chinese Universities
基金
国家自然科学基金(51176125)
关键词
换热网络综合
混沌蚁群算法
局部搜索策略
Powell法
heat exchanger network synthesis
chaotic ant swarm algorithm
local search strategy
Powell method
