一种基于群代理协同交互的共乘出行仿真模型  

CARPOOLING SIMULATION MODEL BASED ON MULTI-AGENT COOPERATIVE INTERACTION

作　　者：唐蕾[1] 孙晨 段宗涛[1] 陈柘[1] Tang Lei;Sun Chen;Duan Zongtao;Chen Zhe(School of Information Technology,Chang’an University,Xi’an 710064,Shaanxi,China)

机构地区：[1]长安大学信息工程学院

出　　处：《计算机应用与软件》2019年第7期83-88,121,共7页Computer Applications and Software

基　　金：陕西省重点科技创新团队项目(2017KCT-29);陕西省重点研发计划项目(2017GY-072)

摘　　要：作为一种新兴的交通方式,共乘出行既环保又可持续,因为它不仅能让乘客节省出行成本,还能减少汽车的交通拥堵,所以共乘成为人们首选的出行方式。考虑到现实中人力资源的有限性,对共乘出行服务进行了深入的研究。对多Agent技术和共乘过程进行详细的论述,提出一种基于群代理协同交互的仿真模型,该模型采用基于三层架构的JADE平台。给出司乘人员路径匹配模型和协商模型,提出最短路径算法来仿真启动司乘代理之间的共乘过程。选取覆盖北京市区的真实轨迹数据集Geolife和西安市15天的出租车营运轨迹数据,对提出的路径匹配模型和协商模型进行实验验证。实验结果表明,两类模型在准确性和执行效率上具有高效性。As a new mode of transportation,carpooling is both environmentally friendly and sustainable,because it can not only save passengers travel costs,but also reduce traffic congestion of cars,so carpooling has become the preferred mode of travel. Considering the limitation of human resources in reality,this paper made an in-depth study on the carpooling service. We discussed multi-agent technology and carpooling process in detail,and proposed a simulation model based on cooperative interaction of group agents,which adopted JADE platform based on three-tier architecture. On this basis,this paper presented the passenger path matching model and negotiation model,and proposed the shortest path algorithm to simulate the carpooling process between passenger agents. We selected Geolife,a real trajectory data set covering Beijing and Xi’an s 15-day taxi operation trajectory data,and tested the proposed path matching model and negotiation model. Experiments show that the two models proposed in this paper are efficient in accuracy and execution efficiency.

关 键 词：共乘出行 多AGENT技术 JADE 路径匹配模型 协商模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]