Publication

Enhancing model-based feedback perimeter control with data-driven online adaptive optimization

Nikolaos Geroliminis, Anastasios Kouvelas, Mohammadreza Saeedmanesh
2017
Journal Articles

Résumé

Most feedback perimeter control approaches that are based on the Macroscopic Fundamental Diagram (MFD) and are tested in detailed network structures restrict inflow from the external boundary of the network. Although such a measure is beneficial for the network performance, it creates virtual queues that do not interact with the rest of the traffic and assumes small unrestricted flow (i.e. almost zero disturbance). In reality, these queues can have a negative impact to traffic conditions upstream of the protected network that is not modelled. In this work an adaptive optimization scheme for perimeter control of heterogeneous transportation networks is developed and the aforementioned boundary control limitation is dropped. A nonlinear model is introduced that describes the evolution of the multi-region system over time, assuming the existence of well-defined MFDs. Multiple linear approximations of the model (for different set-points) are used for designing optimal multivariable integral feedback regulators. Since the resulting regulators are derived from approximations of the nonlinear dynamics, they are further enhanced in real-time with online learning/adaptive optimization, according to performance measurements. An iterative data-driven technique is integrated with the model-based design and its objective is to optimize the gain matrices and set-points of the multivariable perimeter controller based on real-time observations. The efficiency of the derived multi-boundary control scheme is tested in microsimulation for a large urban network with more than 1500 roads that is partitioned in multiple regions. The proposed control scheme is demonstrated to achieve a better distribution of congestion (by creating "artificial" inter-regional queues), thus preventing the network degradation and improving total delay and outflow. (C) 2016 Elsevier Ltd. All rights reserved.

Source officielle

https://infoscience.epfl.ch/entities/publication/71684f8e-58d0-419f-b443-fc26e3c7f981

Nikolaos Geroliminis, Anastasios Kouvelas, Mohammadreza Saeedmanesh
2017
Journal Articles

Résumé

Source officielle

https://infoscience.epfl.ch/entities/publication/71684f8e-58d0-419f-b443-fc26e3c7f981

Concepts associés (25)

Traffic flow

In mathematics and transportation engineering, traffic flow is the study of interactions between travellers (including pedestrians, cyclists, drivers, and their vehicles) and infrastructure (including highways, signage, and traffic control devices), with the aim of understanding and developing an optimal transport network with efficient movement of traffic and minimal traffic congestion problems.

Théorie des trois phases du trafic

La théorie des trois phases du trafic est une théorie alternative de la modélisation du trafic routier mise au point par Boris Kerner entre 1996 et 2002. Elle se concentre principalement sur l'explication physique de la dégradation des conditions de trafic et des embouteillages résultant sur les autoroutes. À la différence des théories classiques basées sur le diagramme fondamental du trafic qui distinguent deux phases ou régimes (trafic fluide et les congestions), la théorie de Kerner propose trois phases.

Algorithme du gradient stochastique

L'algorithme du gradient stochastique est une méthode de descente de gradient (itérative) utilisée pour la minimisation d'une fonction objectif qui est écrite comme une somme de fonctions différentiables. À la fois l'estimation statistique et l'apprentissage automatique s'intéressent au problème de la minimisation d'une fonction objectif qui a la forme d'une somme : où le paramètre qui minimise doit être estimé. Chacune des fonctions est généralement associée avec la -ème observation de l'ensemble des données (utilisées pour l'apprentissage).

Afficher plus

Publications associées (29)

An Iterative Adaptive Dynamic Programming Approach for Macroscopic Fundamental Diagram-Based Perimeter Control and Route Guidance

Nikolaos Geroliminis, Can Chen

Macroscopic fundamental diagrams (MFDs) have been widely adopted to model the traffic flow of large-scale urban networks. Coupling perimeter control and regional route guidance (PCRG) is a promising strategy to decrease congestion heterogeneity and reduce ...

Informs2024

Scalable constrained optimization

Maria-Luiza Vladarean

Modern optimization is tasked with handling applications of increasingly large scale, chiefly due to the massive amounts of widely available data and the ever-growing reach of Machine Learning. Consequently, this area of research is under steady pressure t ...

EPFL2024

On the Generalization of Stochastic Gradient Descent with Momentum

Volkan Cevher, Kimon Antonakopoulos

While momentum-based accelerated variants of stochastic gradient descent (SGD) are widely used when training machine learning models, there is little theoretical understanding on the generalization error of such methods. In this work, we first show that th ...

Microtome Publishing2024

Afficher plus