俞叶锋,张宸,高展,等.基于多目标遗传算法的行驶工况构建优化[J].内燃机工程,2023,44(5):57-65.
基于多目标遗传算法的行驶工况构建优化
Optimization of Driving Cycle Development Based on Multi-Objective Genetic Algorithm
DOI：10.13949/j.cnki.nrjgc.2023.05.008
关键词:行驶工况  多目标优化  遗传算法  非支配排序遗传算法  短行程
Key Words:driving cycle  multi-objective optimization  genetic algorithm（GA）  non-dominated sorting genetic algorithm Ⅱ（NSGA-Ⅱ）  micro-trip
基金项目:国家自然科学基金项目（52071216，52106175）
作者单位E-mail
俞叶锋* 上海交通大学 中英国际低碳学院 leafyu@sjtu.edu.cn 
张宸 上海交通大学 国家电投智慧能源创新学院 chenzhang87@sjtu.edu.cn 
高展 上海交通大学 中英国际低碳学院
上海交通大学 机械与动力工程学院 		gaozhan@sjtu.edu.cn 
伍晨波 上海交通大学 机械与动力工程学院
中国汽车工程研究院股份有限公司，重庆401122 		wuchenbo@caeri.com.cn 
朱磊* 上海交通大学 机械与动力工程学院 tonyzhulei@sjtu.edu.cn 
黄震 上海交通大学 国家电投智慧能源创新学院
上海交通大学 机械与动力工程学院 		z-huang@sjtu.edu.cn 
摘要点击次数: 1140
全文下载次数: 714
摘要:在利用短行程法构建行驶工况的过程中，为综合考虑短行程的典型性和行驶工况的代表性，将二者量化并作为多目标优化过程中的两个目标函数，引入非支配排序遗传算法Ⅱ（non-dominated sorting genetic algorithm Ⅱ， NSGA-Ⅱ）对短行程的选择进行优化。将NSGA-Ⅱ优化方法与常用的短行程选择方法进行对比，在相同时间内，随机选择法、相关性法、距离法和NSGA-Ⅱ分别构建了7 582、7 209、9 615和20个候选行驶工况。结果显示，4种方法中，在整体上，NSGA-Ⅱ产生的行驶工况集中在代表性较高的区域同时使得构成行驶工况的短行程整体上典型性较高；在最优工况的比较中，NSGA-Ⅱ的最优工况与原始数据的特征参数相对误差最小，包含的短行程到离簇中心的平均距离最短，表明采用NSGA-Ⅱ算法对短行程选择过程进行优化有助于从多个角度提升行驶工况的构建质量。
Abstract:In the construction of driving cycles using micro-trip method， to comprehensively consider the typicality of micro-trips and the representativeness of driving cycles， the typicality and representatives were quantified and used as two objective functions in the multi-objective optimization process， and the non-dominated sorting genetic algorithm Ⅱ（NSGA-Ⅱ） was introduced to optimize the selection of micro-trips. The NSGA-Ⅱ optimization method was compared with the commonly used micro-trip selection method， and 7 582， 7 209， 9 615 and 20 candidate driving cycles were constructed by the random selection method， correlation method， distance method and NSGA-Ⅱ， respectively， in the same time. The results show that， among the four methods， on the whole， the driving cycles generated by NSGA-Ⅱ are concentrated in areas with high representativeness while making the micro-trips constituting the driving cycles more typical in general. In the comparison of the optimal cycles， the optimal cycles of NSGA-Ⅱ have the smallest relative errors in the characteristic parameters with the original data and contain the micro-trips to the closest average distance from the cluster center， indicating that the optimization of micro-trip selection process using the NSGA-Ⅱ algorithm helps to improve the quality of the construction of driving cycles from multiple perspectives.
查看全文  HTML   查看/发表评论