| 基于动态拓扑结构的分布式Memetic差分进化算法同步综合换热网络 |
| |
| 引用本文： | 上海理工大学 新能源科学与工程研究所,上海. 基于动态拓扑结构的分布式Memetic差分进化算法同步综合换热网络[J]. 热能动力工程, 2017, 32(2): 34-41 |
| |
| 作者姓名： | 上海理工大学 新能源科学与工程研究所 上海 |
| |
| 作者单位： | 针对标准DE(差分进化)算法在优化换热网络时出现的局部搜索能力弱、易陷入局部最优等问题，本文建立了一种基于动态拓扑结构的分布式Memetic差分进化算法，同步综合换热网络。首先，在子种群内部采用基于欧拉距离的动态拓扑结构，子种群之间采用冯诺依曼拓扑结构，有效地加快了个体之间的信息交流，保持种群多样性，扩大搜索范围。继之，结合Memetic算法思想，将Hooke Jeeves算法作为局部搜索策略，增强算法局部搜索能力。同时，对于局部搜索获得的新解，提出了一种协作学习机制，平衡算法的全局寻优与快速收敛能力。最后，为处理整型变量，提出了两条简单有效的整型变量优化策略，使算法实现了连续变量与整型变量的同步优化。选取两个经典算例验证了算法的有效性。算法应用于算例一，相对于现有文献的最优值，本文所得结构的费用值下降了1 783 $/a，表明算法的性能优于标准DE算法以及其它改进版本的DE算法。算法应用于算例二，相对于现有文献的最优值，本文所得结构的费用值下降了1 209 $/a，表明算法能够有效地处理大规模换热网络问题，具有很强的鲁棒性。 |
| |
| 基金项目： | 国家自然科学基金资助项目(51176125);上海市科委部分地方院校能力建设计划(16060502600) |
| |
| 摘 要： | 针对标准DE(差分进化)算法在优化换热网络时出现的局部搜索能力弱、易陷入局部最优等问题,本文建立了一种基于动态拓扑结构的分布式Memetic差分进化算法,同步综合换热网络。首先,在子种群内部采用基于欧拉距离的动态拓扑结构,子种群之间采用冯诺依曼拓扑结构,有效地加快了个体之间的信息交流,保持种群多样性,扩大搜索范围。继之,结合Memetic算法思想,将Hooke-Jeeves算法作为局部搜索策略,增强算法局部搜索能力。同时,对于局部搜索获得的新解,提出了一种协作学习机制,平衡算法的全局寻优与快速收敛能力。最后,为处理整型变量,提出了两条简单有效的整型变量优化策略,使算法实现了连续变量与整型变量的同步优化。选取两个经典算例验证了算法的有效性。算法应用于算例一,相对于现有文献的最优值,本文所得结构的费用值下降了1 783$/a,表明算法的性能优于标准DE算法以及其它改进版本的DE算法。算法应用于算例二,相对于现有文献的最优值,本文所得结构的费用值下降了1 209$/a,表明算法能够有效地处理大规模换热网络问题,具有很强的鲁棒性。
|
| 关 键 词： | 换热网络综合；差分进化算法；Memetic算法；Hooke-Jeeves算法；整型变量 |
Distributed Memetic Differential Evolution Algorithm based on Dynamic Topology for Simultaneous Synthesis of Heat Exchanger Networks |
| |
| Research Institute of New energy Science and Technology,University of Shanghai for Science and Technology,Shanghai,Chin,Post Code:. Distributed Memetic Differential Evolution Algorithm based on Dynamic Topology for Simultaneous Synthesis of Heat Exchanger Networks[J]. Journal of Engineering for Thermal Energy and Power, 2017, 32(2): 34-41 |
| |
| Authors: | Research Institute of New energy Science Technology University of Shanghai for Science Technology Shanghai Chin Post Code: |
| |
| Affiliation: | Because differential evolution algorithm has weak local search ability and easily converges to a local minimum when optimizing the heat exchanger networks (HEN),a distributedmemetic differential evolution algorithm based on dynamic topology is developed for simultaneous synthesis of HEN. Firstly, subpopulations are connected by the von Neumann topology and individuals in the subpopulation are connected by a dynamic topology based on the Euclidean distance, which can speed up the information exchange and expand the search scope. Then the Hooke-Jeeves algorithm is introduced to improve the local search ability. Meanwhile, for handling the new solution generated by Hooke-Jeeves algorithm, a collaborative learning mechanism is proposed to balance the global search ability and the convergence speed. Finally, two optimizing strategies of integer variables are combined with the proposed algorithm in order to simultaneously optimize continuous and integer variables. Two case studies are selected to demonstrate the validity of this approach. The result of case 1 shows that the proposed algorithm is better than the standard DE algorithm and other improved versions of DE algorithm. The result of case 2 shows that the proposed algorithm can effectively tackle the large scale HEN problem and has a strong robustness. |
| |
| Abstract: | |
| |
| Keywords: | heat exchanger networks synthesis, differential evolution algorithm, Memetic algorithm Hooke-Jeeves algorithm, integer variables |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《热能动力工程》浏览原始摘要信息 |
|
点击此处可从《热能动力工程》下载全文 |
|
