基于实时路况的top-κ载客热门区域推荐
发布时间：2019-02-21 15:41
【摘要】：为降低城市出租车的空载率,缓解路网交通拥堵压力,亟需设计有效的出租车载客热门区域推荐方法.针对传统的出租车相关推荐方法忽略实际路况导致推荐精度较低的现状,提出了一个两阶段的载客热门区域实时推荐算法.首先,离线挖掘阶段,基于出租车历史轨迹数据集提取基于时段属性的载客热门区域;随后,在线推荐阶段,根据出租车请求位置及时间,结合实时路况设计潜在空载时间开销函数T_(cost)对载客热门区域进行评测排序,继而发现Top-κ载客热门区域.基于出租车轨迹数据集的实验结果表明,结合实时交通状况的Top-κ载客热门区域推荐方法以确保较小潜在空载时间开销,相较于传统的出租车推荐方法具有较好的有效性与鲁棒性.
[Abstract]:In order to reduce the no-load rate of urban taxis and relieve the pressure of traffic congestion on the road network, it is urgent to design an effective recommended method for the hot areas where taxis carry passengers. In view of the fact that the traditional taxicab recommendation methods ignore the actual traffic conditions and lead to low recommendation accuracy, a two-stage real-time recommendation algorithm for hot passenger regions is proposed. Firstly, in the stage of off-line mining, the hot area of passenger carrying is extracted based on the historical track data set of taxi. Then, in the online recommendation stage, according to the location and time of taxi request, the potential no-load time cost function T _ (cost) is designed in combination with the real-time traffic condition to evaluate and sort the hot area of passenger carrying, and then the hot area of Top- 魏 is found. The experimental results based on the taxi track data set show that the recommended method of Top- 魏 passenger hot area combined with real time traffic condition can ensure a small potential no-load time cost. Compared with the traditional taxi recommendation method, it has better effectiveness and robustness.
【作者单位】： 西华师范大学计算机学院;中国人民武装警察部队警官学院电子技术系;
【基金】：四川省教育厅重点基金项目(17ZA0381,13ZA0015) 西华师范大学国家培育项目(16C005);西华师范大学英才科研基金(17YC158)
【分类号】：TP391.3;U495
本文编号：2427634
[Abstract]:In order to reduce the no-load rate of urban taxis and relieve the pressure of traffic congestion on the road network, it is urgent to design an effective recommended method for the hot areas where taxis carry passengers. In view of the fact that the traditional taxicab recommendation methods ignore the actual traffic conditions and lead to low recommendation accuracy, a two-stage real-time recommendation algorithm for hot passenger regions is proposed. Firstly, in the stage of off-line mining, the hot area of passenger carrying is extracted based on the historical track data set of taxi. Then, in the online recommendation stage, according to the location and time of taxi request, the potential no-load time cost function T _ (cost) is designed in combination with the real-time traffic condition to evaluate and sort the hot area of passenger carrying, and then the hot area of Top- 魏 is found. The experimental results based on the taxi track data set show that the recommended method of Top- 魏 passenger hot area combined with real time traffic condition can ensure a small potential no-load time cost. Compared with the traditional taxi recommendation method, it has better effectiveness and robustness.
【作者单位】： 西华师范大学计算机学院;中国人民武装警察部队警官学院电子技术系;
【基金】：四川省教育厅重点基金项目(17ZA0381,13ZA0015) 西华师范大学国家培育项目(16C005);西华师范大学英才科研基金(17YC158)
【分类号】：TP391.3;U495
【相似文献】
相关重要报纸文章 前1条
1 通讯员 尚藿;服务TOP—VIP订舱客户[N];中国远洋报;2011年
，本文编号：2427634
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2427634.html
