ABSTRACT

In this paper, continuous timed Petri nets (CTPN) are used to develop a hybrid traffic model, where the network is modelled as a macroscopic model and calibrated by microscopic models. The concept of CTPN is used to build a modular model, where first the highway traffic system is decomposed into several systems, based on structural entities (highway segment, on- and off-ramp links), which are coalesced into a complete model. The result is a light, versatile and easily scalable stochastic model for traffic flow. The calibration and validation of the traffic model is performed through the comparison of basic traffic parameters (flow rate, density, and mean speed) obtained through the traffic model implemented and the commercial micro-modelling software, Aimsun, for part of Portugal’s highway network. The results show that the proposed methodology results in a good trade-off between accuracy, simplicity, and computational cost.

摘要

在本文中，连续定时Petri网（CTPN）用于开发混合交通模型，其中网络被建模为宏观模型并通过微观模型进行校准。CTPN 的概念用于构建模块化模型，其中首先将公路交通系统分解为多个系统，基于结构实体（公路路段、入口和出口匝道链路），这些系统合并为一个完整的模型。结果是一个轻便、通用且易于扩展的交通流随机模型。交通模型的校准和验证是通过比较通过实施的交通模型和商业微建模软件 Aimsun 获得的基本交通参数（流量、密度和平均速度）来进行的，用于葡萄牙的部分高速公路网络。