AOC : Algorithms and Combinatorial Optimisation
Manager : Pierre Fouilhoux
The AOC team was formed in 2010 when the laboratory was restructured into five teams, by some of the members of the OCAD team, bringing together their skills in “Combinatorial Optimisation” and “Algorithms, Software and Distributed Architectures”. The particularity of the AOC team is the variety and the wide spectrum of its research fields and competences, which allow it to work on all theoretical and practical aspects of a problem, and also to elaborate innovative methods from their conception to their efficient implementation.
The research within the team is structured in three main research areas :
- Optimisation in graphs
- Mathematical programming
- Algorithms, software and distributed architectures
Axis 1 : Optimisation in graphs
This line of research focuses on optimisation problems that are modelled in graphs and hypergraphs. We seek to finely characterise the complexity, the solution spaces by polyhedral study, and the degree of approximability. The main results obtained concern: the characterisation of the complexity and approximation of fundamental or applied problems of combinatorial optimisation, the study of the polyhedron induced by the set of feasible solutions of a problem, and finally the exploitation of the properties of graphs and graph algorithms to model and solve in an original way problems that arise in large graphs or specific graphs.
Axis 2: Mathematical programming
This area focuses on the solution (exact or approximate) of combinatorial optimisation problems through mathematical programming. Among the approaches used are constraint and column generation schemes, as well as decomposition and relaxation methods. We also discuss the solution of problems by non-linear models, reoptimisation methods and matheuristic methods.
Area 3: Algorithms, software and distributed architectures
This third axis is federated around distributed computing. Our research is situated at the intersection of the three fields of algorithms, software and distributed architectures. Our research is focused on the following issues: middleware for execution support and scientific workflows in a cloud context, and for massive data management; large-scale system modelling and computation of fault-tolerant properties; parallel and distributed numerical computation; application fault tolerance and irregular parallel applications; placement optimisation for shared and distributed virtual memories and optimisation of cloud services; and service issues.
See the activity report 2012-2017 for more information on the research themes of the AOC team.
AOC seminars : Tuesday at 12.30 pm, manager : Roberto Wolfler Calvo