[DMANET] Ph.D. in LIMOS laboratory with co-supervision in Italy

Dear colleagues,

We are looking for a candidate for a Ph.D. in optimization and artificial
intelligence at the LIMOS laboratory in Clermont-Ferrand, France.

All the details of the offer are presented below.
Thank you for disseminating to your students.

---------------------------------------------------------------------------------
Thesis title: Resolution of inventory and transportation management
problems using Artificial Intelligence.

Location: LIMOS Clermont-Ferrand
Supervisor: Philippe Lacomme (LIMOS)
Co-supervisors: Katyanne Farias (LIMOS), Manuel Iori (DISMI, University of
Modena and Reggio Emilia, Italy)
Start date (desired): as soon as possible (before December 2021)
Application deadline: 1 September 2021
Duration: 3 years

Description of the research subject:
The objective of this thesis is to develop new approaches to solving supply
chain problems over rolling horizons from several weeks to several months.
We aim to create smart planning heuristic methods based on reinforcement
learning techniques, which have been shown to be effective in solving
complex problems of this nature.
This project has as main objective to solve variants of the classical
Inventory Routing Problem (IRP), for which we want to fill some scientific
and technical gaps present in the literature, among which we plan to:
- Include inventory management that explicitly considers the
characteristics of objects (geometry, weight characteristic, costs, etc.)
to obtain management solutions for real applications.
- Add to the IRP the simultaneous management of several products and the
management of a heterogeneous fleet of vehicles that transform the
transportation subproblem into a Heterogeneous Fleet Vehicle Routing
Problem (HVRP).
- Define acceptable solutions by taking into account explicitly and in a
unified manner the notion of quality of service from logistics operators
and customers' point of view.
>From a methodological point of view, the thesis project aims to:
- Characterize the elements that differentiate the good from the bad
solutions by analyzing the structure of the solutions. Then,
reinforcement-based methods will be developed to speed up the convergence.
- Identify the key information allowing to define a local search
procedure taking advantage of global information aiming the treatment of
the problem on long planning horizons.
- Define dynamic adaptation mechanisms for local research movements.
Similar mechanisms were proposed by Thevenin et al. (2019) based on the
Learning Variable Neighborhood Search (LVNS).
- Define mechanisms to analyze a priori the structure of a coding
element, classify it (without carrying out its evaluation) eventually as
not promising, and then avoid costly and unnecessary evaluations.

Required skills:
C/C++ or Java
Notions of Python
Optimization methods (e.g.: metaheuristics, linear programming)
Profile: Bac+5 or more in computer science, optimization, or equivalent.

Contacts :
Philippe Lacomme: placomme@isima.fr
Katyanne Farias: katyanne.farias_de_araujo@sigma-clermont.fr
Manuel Iori: manuel.iori@unimore.it

To apply, contact Ph. Lacomme, K. Farias and M. Iori with a CV and a
recommendation letter.


Best regards,

Katyanne Farias
Maitre de Conferences
SIGMA - LIMOS
Campus de Clermont-Ferrand / Les Cezeaux, CS 20265
63175 Aubiere Cedex, France

**********************************************************
*
* Contributions to be spread via DMANET are submitted to
*
* DMANET@zpr.uni-koeln.de
*
* Replies to a message carried on DMANET should NOT be
* addressed to DMANET but to the original sender. The
* original sender, however, is invited to prepare an
* update of the replies received and to communicate it
* via DMANET.
*
* DISCRETE MATHEMATICS AND ALGORITHMS NETWORK (DMANET)
* http://www.zaik.uni-koeln.de/AFS/publications/dmanet/
*
**********************************************************