Où docteurs et entreprises se rencontrent
Menu
Connexion

Vous avez déjà un compte ?

Nouvel utilisateur ?

Vous aviez déjà un compte sur l’ancien site de l’ABG, afin de le valider aujourd’hui, merci de suivre la procédure de récupération de mot de passe.

Mot de passe oublié

Development of an evolutionary algorithm for calculating multimodal transportation network equilibria

ABG-74410 Stage master 2 / Ingénieur 5 mois < 25 K€ brut annuel
13/02/2018
Logo de
IFSTTAR
Champs sur Marne Ile-de-France France
  • Génie civil, BTP
Optimization, C++, Applied mathematics, Computer science, Operation research
28/02/2017

Établissement recruteur

IFSTTAR is a major player in the European research on the city and the territories, transportation and civil engineering. The French Institute of Science and Technology for Transport, Development and Networks, born on January 1st 2011, from the merger of INRETS and LCPC, is a Public Institution of a Scientific and Technical Nature, under the joint supervision of the Ministry of Environment, Energy and the Sea and the ministry of higher education and research.

Description

5-6 months internship, starting date: March or April 2018

Tutors: Prof. Jean Patrick LEBACQUE (Ifsttar), Mostafa AMELI (Ifsttar)

Location: Ifsttar, Cosys department, Grettia Lab., 14-20 Boulevard Newton, 77420 Champs-sur-Marne

Dynamic transportation systems require proper estimation of travellers’ routing. These problems are called Dynamic Traffic Assignment (DTA) problems. In order to compute the DTA in a large-scale network we use a simulation-based optimization in the solution process. Indeed, several fast and heuristic methods must apply in the optimization process. In the multimodal traffic network, there is no single best solution as equilibrium. Evolutionary algorithms have a successful application to solve complex simulation-based optimization problems. The idea is to apply several types of evolutionary algorithms to solve DTA problems. Moreover, it is important to compare evolutionary algorithms with other common methods.

In this internship, the student, after learning the basics, will work (code C++) in a simulation based optimization platform which uses a traffic simulator for calculating the needed variables in the network. The goal is to apply the evolutionary algorithm (Genetic algorithm and Simulated Annealing algorithm) to the large-scale traffic network and compare it with other methods. By this internship, student will have good knowledge about DTA, traffic simulator and simulation-based optimization.

To this purpose, the steps of this research work are: 


  • To perform a quick bibliographical review of heuristic and evolutionary algorithm used in the traffic assignment context;
  • To study and implement genetic and simulated annealing algorithms for DTA; 

  • To validate the implementation on academic test problems;
  • To apply the algorithm to large-scale network; 

  • To compare the results to the classical optimization implemented in platform; 

  • To study improvements of the proposed algorithm or to suggest a different extension.

Keywords:  User equilibrium, dynamic traffic assignment, large-scale traffic, multimodal, meta-heuristic

References:

Levin, M. W., Pool, M., Owens, T., Juri, N. R., & Waller, S. T. (2015). Improving the convergence of simulation-based dynamic traffic assignment methodologies. Networks and Spatial Economics, 15(3), 655-676.

Duell, M., Amini, N., Chand, S., Grzybowska, H., Saxena, N., & Waller, S. T. (2015, September). Large-scale dynamic traffic assignment: practical lessons from an application in Sydney, Australia. In Intelligent Transportation Systems (ITSC), 2015 IEEE 18th International Conference on (pp. 1735-1740). IEEE.

 

 

Profil

Master 2 student in the field of applied mathematics, computer science, operation research.

Main skills: optimization, skill in computer programming (C++), good level of English, Basic knowledge in evolutionary algorithms, knowledge of transportation and operation research domain will be a plus

Know-how: Autonomy, sense of initiative, excellent relationship, rigor, taste for experiments

Prise de fonction

01/03/2018
Postuler
Fermer

Vous avez déjà un compte ?

Nouvel utilisateur ?


Mot de passe oublié
Ce site est tout nouveau : si vous rencontrez un dysfonctionnement ou si vous avez des suggestions, nous vous serons très reconnaissants de nous en informer. Merci d’avance !
item1 item1
item1