Adaptative Combinatorial Search for E-science

This project finished in 2013.

Project Summary

The goal of this research is to improve the applicability of constraint-based or heuristic-based solvers to complex scientific problems. As demonstrated by a large literature, e-Scientists already benefit from the use of search procedures to tackle a large variety of important problems. Unfortunately, these applications suffer from the limits of current solving technologies which appear to be poorly adapted to these new domains. One solution to improve performance is the fine tuning of the solver parameters. This is a tedious and time-consuming task that often requires knowledge about both the domain and the algorithm.

This approach is hardly applicable to e-Science whose applications fields are constantly growing. We claim that the self-adaptation of a solver to the domain of interest is the only viable solution to this problem. Our goal in this project is to develop tools able to automatically choose the optimal parameter configuration of a given search algorithm for a given problem or class of problems.

This adaptation would allow us to deploy combinatorial search with new complex scientific problems with good expected performance.

Related Stuff