TY - JOUR
T1 - An improved pheromone-based vehicle rerouting system to reduce traffic congestion
AU - Ho, Mun Chon
AU - Lim, Joanne Mun Yee
AU - Soon, Kian Lun
AU - Chong, Chun Yong
N1 - Publisher Copyright:
© 2019 Elsevier B.V.
Copyright:
Copyright 2019 Elsevier B.V., All rights reserved.
PY - 2019/11
Y1 - 2019/11
N2 - The growing number of vehicles necessitates the implementation of effective vehicle rerouting systems. Designing an effective vehicle rerouting system is challenging due to the dynamic nature of vehicular network. In this paper, a Proactive Travel-time based Pheromone Rerouting (PTPR) system is proposed. First, PTPR system predicts future congestion level using travel time and vehicle density information. Then, vehicles are distributed to multiple paths to balance the traffic load. Different from the existing pheromone-based rerouting systems, each ant (vehicle) in PTPR system can deposit its pheromone on multiple road segments away, instead of its direct adjacent road segment, based on its route. This new pheromone model aims to improve the performance of PTPR system. In addition, a localized dynamic k-shortest path (LDkSP) algorithm is proposed to reduce computational effort of PTPR system. Experiments were conducted on two different areas (i.e. suburban and urban) using Simulation of Urban Mobility (SUMO). Results show that the proposed PTPR system outperforms the existing rerouting system by reducing mean travel time, fuel consumption, and increasing number of arrive vehicles by 8.2%, 2%, and 15.1% respectively in Woodlands (suburban) and 28.7%, 17.2%, and 29.5% respectively in Novena (urban). The computation time used to reroute each vehicle is also reduced by 68.3% and 92.1% in suburban and urban area respectively using the proposed LDkSP. Finally, experiments over various usage rates and estimation errors showed that the proposed PTPR system is robust to usage rates ranging from 80% to 100% and is able to function properly with estimation error of up to 20%.
AB - The growing number of vehicles necessitates the implementation of effective vehicle rerouting systems. Designing an effective vehicle rerouting system is challenging due to the dynamic nature of vehicular network. In this paper, a Proactive Travel-time based Pheromone Rerouting (PTPR) system is proposed. First, PTPR system predicts future congestion level using travel time and vehicle density information. Then, vehicles are distributed to multiple paths to balance the traffic load. Different from the existing pheromone-based rerouting systems, each ant (vehicle) in PTPR system can deposit its pheromone on multiple road segments away, instead of its direct adjacent road segment, based on its route. This new pheromone model aims to improve the performance of PTPR system. In addition, a localized dynamic k-shortest path (LDkSP) algorithm is proposed to reduce computational effort of PTPR system. Experiments were conducted on two different areas (i.e. suburban and urban) using Simulation of Urban Mobility (SUMO). Results show that the proposed PTPR system outperforms the existing rerouting system by reducing mean travel time, fuel consumption, and increasing number of arrive vehicles by 8.2%, 2%, and 15.1% respectively in Woodlands (suburban) and 28.7%, 17.2%, and 29.5% respectively in Novena (urban). The computation time used to reroute each vehicle is also reduced by 68.3% and 92.1% in suburban and urban area respectively using the proposed LDkSP. Finally, experiments over various usage rates and estimation errors showed that the proposed PTPR system is robust to usage rates ranging from 80% to 100% and is able to function properly with estimation error of up to 20%.
KW - Long short-term memory
KW - Pheromone
KW - Traffic congestion
KW - Traffic prediction
KW - Travel time
KW - Vehicle routing
UR - http://www.scopus.com/inward/record.url?scp=85070652643&partnerID=8YFLogxK
U2 - 10.1016/j.asoc.2019.105702
DO - 10.1016/j.asoc.2019.105702
M3 - Article
AN - SCOPUS:85070652643
SN - 1568-4946
VL - 84
JO - Applied Soft Computing
JF - Applied Soft Computing
M1 - 105702
ER -