Trajectory Comparison in a Vehicular Network I: Computing a Consensus Trajectory
Trajectory Comparison in a Vehicular Network I: Computing a Consensus Trajectory
复制标题
车载网络中的轨迹比较 I:计算共识轨迹
DOI:
10.1007/978-3-030-23597-0_43
复制
发表时间:
2019
期刊:
影响因子:
--
通讯作者:
Zhu, Binhai
中科院分区:
文献类型:
--
作者:
Zou, Peng;Qingge, Letu;Yang, Qing;Zhu, Binhai
In this paper, we investigate the problem of computing a consensus trajectory of a vehicle giving the history of Points of Interest (POIs) visited by the vehicle over certain period of time. The problem originates from building the social connection between two vehicles in a vehicular network. Formally, given a set ofmtrajectories (sequences’s over a given alphabet, each with length at mostO(n), with), the problem is to compute a target (median) sequenceToversuch that the sum of similarity measure (i.e., number of adjacencies) betweenTand all’s is maximized. For this version, we show that the problem is NP-hard and we present a simple factor-2 approximation. IfThas to be a permutation, then we show that the problem is still NP-hard but the approximation factor can be improved to 1.5. We implement the greedy algorithm and a variation of it which is based on a more natural greedy search. Using simulated data over two months (e.g.,) and variants ofand(e.g.,and), the empirical results are very promising and with the local adjustment algorithm the actual approximation factor is between 1.5 and 1.6 for all the cases.
DOI:
--
发表时间:
1998
期刊:
影响因子:
--
作者:
D. Bryant
通讯作者:
D. Bryant