Date and time : 16/04/2015, 10 am

In our previous works, we have proposed several kinds of approaches that combine knowledge and combinatorial optimization. In this talk, we will first focus on how knowledge discovery tasks can be modelized as combinatorial optimization problems and  then, solved thanks to metaheuristics. We will give some successful applications in health and biology. Secondly, we will present how knowledge can help to improve the search by understanding what makes a solution good/interesting or not for one problem. Finally, we will discuss about how knowledge on fitness landscape can help to design new metaheuristics.