Le 15/11/2020 par elloumi :
Postdoc offer in OR at the Applied Mathematics Unit at ENSTA Paris
Optimized deployment planning of a multi-technology communication network via mobile units
When projecting French forces during external operations, a deployment plan for the communication relays associated with different networks must be established using various technologies: V-UHF, THF, 5G, LDT, Aero Optics, Space Optics, Satcom, .... In this study, the relays can be installed on mobile units. These units have a fixed autonomy and data must regularly be collected from them during the operation.
The problem is to determine simultaneously:
(i) the initial plan for the geographical deployment of mobile units in the field;
(ii) the timing of the data collection of these units over a time horizon;
(iii) for each deployed unit, the list of relays it carries.
The overall solution is evaluated both on technical criteria (coverage, quality of service, network reliability, resistance to jamming, mass/volume/power required per mobile unit, etc.), and on operational criteria such as implementation complexity, cost (financial, HR, ...), etc. The technical criteria must necessarily be satisfied and will therefore represent constraints of the problem. Operational criteria, on the other hand, do not necessarily have a threshold and can be considered as different objectives to optimize.
Finally, since the terrain is uncertain, several approaches of optimization with uncertainty may be considered, depending on the nature of the uncertainty and the data available. For example, if probability distributions are available, a best solution on average would be desired. Otherwise, a robust solution which satisfies most of the scenarios and which can efficiently be adapted to worst cases would be appreciated.
Starting from a more precise description of the problem and realistic data, the objective is to perform a study through Mathematical Programming: MILP modeling and practical resolution as well as integration of the multi-objective and uncertain nature of the problem. An important issue is to obtain results on the proposed instances and to demonstrate the efficiency of the approach and its scaling up.
Ph.D. in Operational Research with discrete mathematical programming skills. Knowledge of multi-objective optimization or in handling uncertainty would be appreciated. The candidate must be of European nationality.
Start and duration
From January 2021, duration 1 year, renewable
Unité de Mathématiques Appliquées, Ensta Paris, 828, Boulevard des Maréchaux, 91762 Palaiseau
Contacts and application
Please send a CV, a short motivation letter and contact information of two references to Sourour Elloumi (email@example.com) and Zacharie Ales (firstname.lastname@example.org)