Big Optimization aNd Ultra Scale computing
- July 1, 2017 — Creation of the BONUS team
- January 30, 2018 — Website is up
- September 27, 2018 — Relocation of BONUS at EuraTechnologies (Place building)
The "Big Optimization aNd Ultra Scale computing" (Bonus) research group is a joint team between the University of Lille, CRIStAL research center (UMR 9189, Univ Lille, CNRS, EC Lille) and the Inria Lille - Nord Europe research center, France.
Big optimization problems (BOPs) refer to problems composed of a large number of environmental input parameters and/or decision variables (high dimensionality), and/or many objective functions that may be computationally expensive. Solving BOPs raises at least four major challenges: (1) tackling their high (curse of) dimensionality ; (2) handling many objectives ; (3) dealing with computationally expensive objective functions ; and (4) scaling on (ultra-scale) modern supercomputers.
The overall scientific objective of the Bonus team is to address efficiently these challenges using the three following research lines:
- Decomposition-based optimization. Given the large scale of the targeted optimization problems of Bonus in terms of the number of variables and objectives, their decomposition into smaller, easier to solve and loosely coupled or independent subproblems is essential to raise the challenge of scalability.
- Machine Learning-assisted Optimization. The objective of ML-aided optimization is to raise the challenge of expensive functions of Big Optimization problems (BOPs) using surrogates but also to assist the two other research lines in dealing with the other challenges of Bonus (high dimensionality and scalability).
- Ultra-scale optimization. This research line intensifies our difference from other (project-)teams of the related Inria scientific theme. It is complementary to the two other ones, which are sources of massive parallelism and with which it is combined to solve BOPs. Indeed, ultra-scale computing is necessary for the effective resolution of the large amount of subproblems generated by decomposition of BOPs, the parallel evaluation of simulation-based fitnesses and metamodels, etc.