刘刚^{1, 2}，孙佳琦^{2, 3}，董伟星⁴
AuthorsHTML:刘刚^{1, 2}，孙佳琦^{2, 3}，董伟星⁴
AuthorsListE:Liu Gang^{1, 2}，Sun Jiaqi^{2, 3}，Dong Weixing⁴
AuthorsHTMLE:Liu Gang^{1, 2}，Sun Jiaqi^{2, 3}，Dong Weixing⁴
Unit:1. 天津大学建筑学院，天津 300072；
2. 天津市建筑物理环境与生态技术重点实验室，天津 300072；
3. 天津大学国际工程师学院，天津 300072；
4. 中国中建设计集团有限公司，北京 100037
Unit_EngLish:1. School of Architecture，Tianjin University，Tianjin 300072，China；
2. Tianjin Key Laboratory of Architectural Physical Environment and Ecological Technologies，Tianjin 300072，China；
3. Tianjin International Engineering Institute，Tianjin University，Tianjin 300072，China；
4. China Construction Engineering Design Group Corporation Limited，Beijing 100037，China
Abstract_Chinese:粒子群及其改进算法是进行建筑能耗优化的重要方法，但是算法性能很大程度上取决于其参数设置．目前针对改进粒子群算法在建筑性能优化领域的最优参数设置的研究较少．本文旨在探讨两种常见的改进粒子群算法：差分粒子群(dPSO)算法和遗传粒子群(gPSO)算法在建筑能耗优化中的最优参数设置问题．在使用测试函数验证改进算法的有效性后，针对以能耗为目标的办公建筑形体优化问题，使用15组常见的参数组合进行重复实验．建立以稳定性、准确性和收敛时间3个指标为目标的算法性能多目标评价模型，计算pareto解集，得到性能表现优异的算法参数组合，即进行建筑能耗优化时，当对计算速度或计算准确度有较高要求时，建议采用参数设置为c1＝c2＝1.5，pm＝0.5，pc＝0.9 或c1＝c2＝2.0，pm＝0.1，pc＝0.9 的gPSO算法；当对优化过程没有偏好时，可采用参数设置为c1＝c2＝2.0，CR＝0.5，F＝0.4 的dPSO算法．最后使用不同气候区的同类型建筑优化问题，对得到的高效参数组合进行了验证．
Abstract_English:Particle swarm optimization and its improved algorithms are important methods of building performance optimization. However，the performance of the evolutionary algorithm during optimization is largely dependent on its parameter settings. At present，only a few studies of the optimal parameter settings of the improved particle swarm optimization algorithm in the field of building performance optimization have been conducted. In this study，the optimal parameter settings of two commonly used improved particle swarm optimization algorithms，i.e.，differential particle swarm optimization(dPSO)algorithm and genetic particle swarm optimization(gPSO)algorithm，are discussed to solve the building energy consumption optimization problem. After the test functions were utilized to verify the effectiveness of the improved algorithm，15groups of common parameter combinations were used to perform repeated experiments on the optimization problem of office buildings，with energy consumption as the target. On the basis of three evaluation indices，i.e，stability，accuracy，and convergence time，a multi-objective model for the evaluation of the performance of the algorithm for the building optimization problem was established，pareto solution set was calculated，the parameter combination with excellent performance was determined：when computational speed or accuracy are required for the building energy consumption optimization process，gPSO with the parameter settings c1＝c2＝1.5，pm＝0.5，pc＝0.9 or c1＝c2＝2.0，pm＝0.1，pc＝0.9can be adopted. When neither computational speed nor accuracy is required for the optimization process，dPSO with the parameter setting c1＝c2＝2.0，CR＝0.5，F＝0.4can be adopted. Finally，the same type of building optimization problem in different climate zones was used to verify the efficiency of the parameter combination.
Keyword_Chinese:建筑能耗优化；粒子群优化；差分算子；遗传算子；参数设置
Keywords_English:building energy consumption optimization；particle swarm optimization；differential operator；genetic operator；parameter setting

PDF全文下载地址:http://xbzrb.tju.edu.cn/#/digest?ArticleID=6578