Network signal setting design with stage sequence optimisation

doi:10.1016/j.trb.2017.01.013

2017-06-01

PRE

Giulio Erberto Cantarella *

Stefano De Luca

Roberta Di Pace

DOI：10.1016/j.trb.2017.01.013

原文链接

原文求助

摘要

En 中文

One of the most straightforward short term policies to mitigate urban traffic congestion is control through traffic lights at a single junction or network level. Existing approaches for single junction Signal Setting Design (SSD) can be grouped into two classes: Stage-based or Phase-based methods. Both these approaches take the lane marking layouts as exogenous inputs, but lane-based optimisation method may be found in literature, even though for isolated signal-controlled junctions only. The Network Signal Setting Design (NSSD) requires that offsets are introduced; a traffic flow model is also needed to compute total delay. All existing methods for NSSD follow a stage-based approach; these methods do not allow for stage matrix optimisation: it is shown that explicit enumeration of stage sequences is only practicable for very small networks. This paper focuses on Network Signal Setting Design introducing the so-called scheduled synchronisation that includes green scheduling, green timing and coordination into one optimisation problem. The paper proposes a stage-based method to solve such a problem, as an extension of the synchronisation method and the traffic flow model proposed in Cantarella et al. (2015): first a set of candidate stages is defined for each junction, then the stage sequences, the stage lengths and the offsets are optimised all together. To the authors' knowledge, no other one-step optimisation method is available in literature for scheduled synchronisation. Results of the proposed method to a small network were compared with those from explicit enumeration of all stage sequences; results for a larger network are also discussed. (C) 2017 Elsevier Ltd. All rights reserved.

Keyword:

Network signal setting design

Scheduled synchronisation

Stage based methods

Macroscopic traffic flow modelling

Meta-heuristics