蚁群算法在道路应急疏散策略选择中的应用  

Application of ACO algorithm to road emergency evacuation strategy selection

作　　者：毛新华[1] 王建伟[1] 赵京[1] 甘家华[1] 

机构地区：[1]长安大学经济与管理学院,陕西西安710064

出　　处：《中国安全科学学报》2014年第10期170-176,共7页China Safety Science Journal

基　　金：国家自然科学基金资助(51278057);国家社会科学基金资助(09XJY004);陕西省自然科学基金资助(2012JQ5013)

摘　　要：为有效节约道路突发事件发生后车辆的应急疏散时间,基于疏散车辆交通流和非疏散交通流的不确定性,建立带有区间参数的机会约束规划模型,研究道路应急安全疏散策略的选择问题。以某疏散路网为例,模拟道路突发事件发生后的车辆疏散。用蚁群算法(ACO)计算最优和最差策略。结果表明:疏散路网饱和度上限和置信水平的设置影响疏散策略选择和疏散时间,置信水平取值越高,总疏散时间越短;由于路网非疏散流概率密度的影响,路段饱和度上限取值与整体疏散时间不存在明显的正相关关系。当置信水平取0.95、饱和度上限取0.9时,将得到最优疏散策略;当置信水平取0.8、饱和度上限取0.75时,将得到最差疏散策略。To save the evacuation time after a road emergency,a chance constrained programming model with interval parameters was built based on uncertainty of traffic flows of both evacuation vehicles and nonevacuation vehicles. The problem of road emergency evacuation strategy selection was studied by using the model. Vehicle evacuation was simulated in a certain road network with three exits. Vehicles＇ optimal and worst evacuation strategies were calculated by ACO algorithm. The calculating results show that evacuation strategy selection and evacuation time are influenced by upper limit of evacuation network saturation and confidence level setting,and higher value confidence level is set,the shorter the total evacuation time will be,but with influence of non-evacuation flow probability density,there is no obvious positive correlation relationship between upper limit of evacuation network saturation and total evacuation time,that when confidence level is 0. 95,and upper limit of evacuation network saturation is 0. 9,optimal evacuation strategy can be obtained,and that when confidence level is 0. 8,and upper limit of evacuation network saturation is 0. 75,worst evacuation strategy can be obtained.

关 键 词：交通工程 疏散策略选择 机会约束规划 不确定性 蚁群算法(ACO) 

分 类 号：X913.4[环境科学与工程—安全科学] U491[交通运输工程—交通运输规划与管理]