带混合时间窗的多目标危险化学品运输路径优化  

Route Optimization of Hazardous Chemicals Transportation with Mixed Time Windows

作　　者：邹宗峰[1,2] 张保全[1] 

机构地区：[1]上海大学管理学院,上海200444 [2]复旦大学管理学院,上海200433

出　　处：《中国安全科学学报》2012年第4期83-89,共7页China Safety Science Journal

基　　金：上海市科委软科学项目(11692110002);上海市哲学与社会科学规划项目(2011BGL008);教育部人文社会科学研究项目(11YJC630309)

摘　　要：为降低危险化学品运输风险,使路径选择更加科学、切合实际,从责任主体、社会影响等方面综合考虑危险化学品运输相关主体,根据各主体目标对影响运输路径选择的指标进行重新识别和细分,提出5大指标体系。应用熵权法计算出各指标权重,建立混合时间窗条件下的多目标危化品运输路径优化模型,通过改进的多目标遗传算法求出Pareto最优解。结果表明,不同时间点出发会产生不同的Pareto最优解,并且可得到每条最优路径的出发时间窗,每个最优解代表了某项指标达到相对最优时的路径。混合时间窗较符合实际情况,基于出发时间窗及生成的多条最优路径,决策者可根据决策偏好及关注重点以较高效率选出最合适的路径。For the sake of reducing the risk of dangerous chemicals transportation,and making path selection more scientific and in line with the reality,different relevant stakeholders involved in the transportation were considered from different aspects（the main body of responsibility,social impact and so on）,and five indicator systems were proposed on the basis of the re-identification and segmentation of indexes that affect the transport routes chosen seriously.The weights of these indexes were calculated by using the entropy weight method.The multi-objective models for hazardous chemicals transport were built on the basis of the mentioned five indicators under the condition of mixed time windows.The Pareto optimal solutions were obtained by using the improved multi-objective genetic algorithm.Analysis shows that different starting time point of departure will produce different Pareto optimal solutions,which including departure time window of each optimal path,and each optimal solution represents a path with an indicator having reached relatively optimal.It is concluded that the conditions of mixed time windows are more in line with the actual situation.On the basis of the departure time window and the generated multiple optimal paths,decision-makers can choose the most appropriate optimal path with higher efficiency according to their decision-making preferences and their focus of attention.

关 键 词：危险化学品运输 路径优化 混合时间窗 多目标遗传算法 熵权法 

分 类 号：X937[环境科学与工程—安全科学]