Post-Doctoral - optimization - Orange Labs (France)
Forum 'Emplois' - Sujet créé le 2008-06-05
Job Title: Post-Doctoral position in GSM capacity Optimization in UMTS900 rollout context
Profile: Post Doctoral
Duration: 12 months (+ 6 renewable under conditions)
Scope: Combinatorial Optimization, Mathematical Modelling
Place: Orange Labs' premises in Belfort (East of France)
Date Posted: 01/06/2008
Validity: 4 months
To apply: send statement of research interests & your resume at recruit_ti889@rd.orange-ftgroup.com
Job description: [/b]
In GSM networks, due to interference, spectrum management is necessary to ensure capacity and service quality. Because of the size of real networks, still growing, and to the restricted number of frequencies in the available radio spectrum, the frequency assignment appears as a very competitive challenge to deal with.
The "classical" Frequency Assignment Problem (FAP) may be understood as an optimization problem where the goal is to minimize co-channel and adjacent-channel interference subject to a various set of constraints to satisfy. The challenge consists in assigning a predefined number of frequency channels to each transmitter of a sector to satisfy traffic from users on its coverage area. The large size of real networks (up to several thousands of sectors) and spectrum scarcity (down to 40 frequency channels per operator for 900 MHz GSM) involve a high rate of frequency reuse.
Opening of UMTS service in a shared bandwidth (900MHz) with current GSM networks decreases dramatically the number of available frequency channels for GSM frequency planning (between 21 and 25 channels less).
It leads frequency planning to a new question which is: how many transmitters (and which transmitters) does an operator have to suppress on its GSM network in order to keep a quality to the situation before the opening of UMTS900 service? Or in another words, what is the loss of GSM capacity that has to be taken over by other layers of the network to keep and acceptable interference quality of its GSM frequency plan?
This new optimization problem will have to be analyzed, and modeled within current optimization modeling in Orange Labs frequency planning software.
A combinatorial optimization solution will have to be proposed, implemented and tested in order to efficiently solve this new problem.
Outcomes will be validated on real networks.
Desired skills:[b]
The position requires knowledge in combinatorial problem solving method, experience in modeling (mathematics) and in development (C++ code). Background in wireless radio mobile technology is desirable.
The applicant is expected to be successful in both working independently and collaboratively with researchers.
Profile: Post Doctoral
Duration: 12 months (+ 6 renewable under conditions)
Scope: Combinatorial Optimization, Mathematical Modelling
Place: Orange Labs' premises in Belfort (East of France)
Date Posted: 01/06/2008
Validity: 4 months
To apply: send statement of research interests & your resume at recruit_ti889@rd.orange-ftgroup.com
Job description: [/b]
In GSM networks, due to interference, spectrum management is necessary to ensure capacity and service quality. Because of the size of real networks, still growing, and to the restricted number of frequencies in the available radio spectrum, the frequency assignment appears as a very competitive challenge to deal with.
The "classical" Frequency Assignment Problem (FAP) may be understood as an optimization problem where the goal is to minimize co-channel and adjacent-channel interference subject to a various set of constraints to satisfy. The challenge consists in assigning a predefined number of frequency channels to each transmitter of a sector to satisfy traffic from users on its coverage area. The large size of real networks (up to several thousands of sectors) and spectrum scarcity (down to 40 frequency channels per operator for 900 MHz GSM) involve a high rate of frequency reuse.
Opening of UMTS service in a shared bandwidth (900MHz) with current GSM networks decreases dramatically the number of available frequency channels for GSM frequency planning (between 21 and 25 channels less).
It leads frequency planning to a new question which is: how many transmitters (and which transmitters) does an operator have to suppress on its GSM network in order to keep a quality to the situation before the opening of UMTS900 service? Or in another words, what is the loss of GSM capacity that has to be taken over by other layers of the network to keep and acceptable interference quality of its GSM frequency plan?
This new optimization problem will have to be analyzed, and modeled within current optimization modeling in Orange Labs frequency planning software.
A combinatorial optimization solution will have to be proposed, implemented and tested in order to efficiently solve this new problem.
Outcomes will be validated on real networks.
Desired skills:[b]
The position requires knowledge in combinatorial problem solving method, experience in modeling (mathematics) and in development (C++ code). Background in wireless radio mobile technology is desirable.
The applicant is expected to be successful in both working independently and collaboratively with researchers.