Relieve the congestion by shuttle bus in rush hours using aggregation clustering algorithm on group travel pattern

摘要

The travel demand is of vital importance for transport planning. Especially in rush hours, the commutersaggregate in certain areas, resulting in traffic jams, which hinder the normal operation ofthe urban traffic. The key issue to solve the problem is to find out the demand of the passengersand accordingly arrange the transit resource more reasonably. This paper proposes a frameworkthat provides a shuttle bus solution to satisfy the travel demand of the gathered passengers inrush hours according to their travel history obtained by smartcard data. Firstly, an aggregationalgorithm basing on Clustering by fast search and find of density peaks (CFSFDP) is presented tohighlight the area, which consists of adjacent bus stations with high passenger flow, addressed asspark region. Secondly, group travel pattern (GTP) is put forward to describe the travel trend on acity-wide scale, which reveals the common travel demand of the commuters. Lastly, an algorithmnamed Variable Visibility Path Optimization Algorithm based on ant colony algorithm is proposed tomake schedule solution of shuttle bus according to GTP basing on the historical running informationof the bus. The experiment basing on the bus passenger flow data collected by AFCS in Aug2014 from Beijing shows that our method helps to ease the traffic aggregation effectively andpractically and offer reference to the bus scheduling issue.