ABSTRACT

In this paper, a network-wide anticipatory control (AC) framework, incorporating drivers' route choice behaviour, is proposed. The proposed AC, consisting of two main levels of control and route choice, explicitly accounts for road users' responses to the implemented control. Perimeter control, modelled in an MPC framework for real-time applications, optimizes inter-regional transferring flow at the network level. User equilibrium is established through a logit route choice model and solved using the method of successive averages (MSA). The modelling approach adopted is based on macroscopic fundamental diagram (MFD), which provides a unimodal low-scatter relationship between density and outflow of any region within a network. Numerical results indicate that the proposed AC outperforms no control and basic control cases and shows promising results in alleviating congestion and deriving the network to near-system optimum traffic condition, under various demand patterns and levels.

摘要

本文提出了一种包含驾驶员路线选择行为的全网络预期控制（AC）框架。拟议的AC由控制和路线选择两个主要级别组成，明确说明了道路使用者对已实施控制的响应。在MPC框架中为实时应用建模的边界控制在网络级别优化了区域间的传输流。用户平衡通过对数路径选择模型建立，并使用连续平均法（MSA）求解。所采用的建模方法基于宏观基本图（MFD），它在网络内任何区域的密度和流出量之间提供了单峰低散射关系。