Sujet de thèse de doctorat en transport en commun à la demande à l'Université Gustave Eiffel

Forum 'Emplois' - Sujet créé le 04/02/2020 par hosteins (739 vues)

Le 04/02/2020 par hosteins :

Ci-dessous le détail (en anglais) d'un sujet de thèse de doctorat proposée sur le site de Lille de l'Université Gustave Eiffel:


PhD thesis: « Optimised management of a fleet of electric autonomous
vehicles »

Research fields : Computer Science, Operational Research, Artificial Intelligence
Key words : Autonomous vehicle, electric vehicle, optimisation, dynamic algorithms, on-demand transportation

Context and objectives:
The latest reports by environmental agencies show that our mobility system is unsustainable: emissions in the transportation sector have increased by 10% between 1990 and 2018 according to the French High Counsel for Climate. A possible action to reduce these emissions is the improvement of the performances of collective transportation systems.
For example, there is an increasing need for a more flexible Public Transport (PT) system, allowing the service of low density suburban areas at a limited cost, as well as a growing preoccupation concerning polluting emissions of urban and suburban PT. A possible answer to these challenges is the deployment of on-demand public transport services based on a fleet of electric vehicles (Archetti et al 2018). In order to guarantee the necessary level of flexibility for on-demand transportation and minimise costs, the use of autonomous vehicles is often suggested. Several approaches exist to optimally handle on-demand transportation in different contexts, see, e.g., the
so-called Dial A Ride Problem (DARP) (Ho et al 2018). These approaches often apply to very specific situations. A big part of the literature focuses on a taxi configuration where only one passenger is present in a given vehicle at all times (Hyland and Mahmassani 2018), or on cases with very specific requirements such as the ones linked to the mobility of disabled people (Lehuédé et al
The PhD thesis objective is the design of optimisation algorithms for the planning of an on-demand transport system exploiting autonomous electric vehicles in urban and suburban PT. The infrastructure on which the vehicles will travel will be a dedicated infrastructure, with only a few intersections with regular traffic. It will be composed of a main central corridor from which
secondary branches will depart. On the central corridor, the vehicles will follow a regular schedule while on the secondary branches they will operate an on-demand service. Therefore, the problem will be fundamentally different from on-demand transportation problems studied up to now, whose main issue is usually to build the itinerary between the pickup and delivery points, without worrying
about inserting said itinerary in an existing service. It will, however, be closer to the problems of railway traffic management for which the COSYS-ESTAS and COSYS-LEOST laboratories of the Gustave Eiffel University display a long dated expertise (Pellegrini et al 2014, 2015, 2016).
The demand on secondary branches will be considered as provided through a mobile application: the user will indicate the initial and final stations of his trip and the preferred pickup time. The optimisation tool designed during the thesis will decide which vehicle will handle each demand and in which order, all the while respecting the scheduled service on the central infrastructure. The
objective of the optimisation algorithm will be to maximise the level of service (for example minimising the waiting time or the travel time of the passengers). The algorithm will have to provide a solution in real time: the problem is thus intrinsically dynamic and will necessitate frequent reoptimisations. Such problems usually call for the use of heuristic or metaheuristic algorithms to obtain good quality solutions in a limited amount of time. The algorithms will be tested on nominal as well as perturbed scenarios. The perturbations will have different possible causes, e.g., the bad behaviour of some users (e.g. late arrival with respect to the agreed pickup time
or door blockage on a vehicle), a higher flow of passengers than anticipated, hardware failures or unexpected changes of traffic light cycled at intersections along the route.
The autonomous characteristic of the vehicles implies the absence of constraints linked to theworking time of the driver, while their electric propulsion imposes constraints in terms of residual charge and charging time.
After the design, implementation and assessment of the algorithms, the system proposed will be compared with a system of High Level Service Buses (HLSB), inspired by those already working in different cities, that might cover the same territory. This comparison will allow the identification of conditions in which the autonomous vehicles might be the most advantageous PT system from
different points of view (service provider, user and the environment), as HLSB are usually designed to serve mainly dense areas.

PhD advisors:
Paola Pellegrini, University Gustave Eiffel, IFSTTAR, COSYS-LEOST :
Pierre-Olivier Vandanjon, University Gustave Eiffel, IFSTTAR, AME-EASE:
Pierre Hosteins, University Gustave Eiffel, IFSTTAR, COSYS-ESTAS:

University Gustave Eiffel, IFSTTAR, COSYS-ESTAS laboratory
Villeneuve d’Ascq (Lille), France

Required skills (one or more):
- Operational Research methods (combinatorial optimisation, Linear Programming, metaheuristics, online optimisation,...)
- software design
- C++ coding
- English spoken and written

Documents required:
- Detailed CV
- Program and grades of the master courses
- Letter of motivation
- Optionally, one or more letters of recommendation, if the candidate thinks it is useful.

Bibliographic references:
P. Pellegrini., G. Marlière, R. Pesenti and J. Rodriguez (2015). RECIFE-MILP: an effective MILP-based heuristic for the real-time railway trafic management problem. IEEE Transactions on Intelligent Transportation Systems, 16(5): 2609-2619.
P. Pellegrini, G. Marlière and J. Rodriguez (2014). Optimal train routing and scheduling for managing traffic perturbations in complex junctions. Transportation Research Part B, 59: 58-80.
M. Samà, P. Pellegrini, A. D’Ariano, J. Rodriguez and D. Pacciarelli (2016). Ant colony optimization for the real-time train routing selection problem. Transportation Research Part B, 85: 89-108.
S.C. Ho, W.Y. Szeto, Y.-H. Kuo, J.M.Y. Leung, M. Petering and T.W.H. Tou (2018). A survey of dial-a-ride problems: Literature review and recent developments. Transportation Research Part B, 111: 395-421.
C. Archetti, M. Grazia Speranza and W. Dennis (2018). A simulation study of on-demand transportation system. International Transactions in Operational Research, 25(4): 1137-1161.
M. Hyland and H.S. Mahmassani (2018). Dynamic autonomous vehicle fleet operations: Optimization-based strategies to assign AVs to immediate traveler demand requests.
Transportation Research Part C, 92: 278-297.F. Lehuédé, R. Masson, S.N. Parragh, O. Péton and F. Tricoire (2014). A multi-criteria large
neighbourhood search for the transportation of disabled people. Journal of the Operational Research Society, 65: 983-1000.

Moteur de recherche
Tous les forums

  La Société française de Recherche Opérationnelle et Aide à la Décision ROADEF est une association Loi 1901 Plus d'informations sur la ROADEF