This project is devoted to the study of Mathematical Programming -- or numerical optimization -- from the theoretical and practical viewpoints, as well as its application to problems of industrial origin. This includes : the theory of optimality conditions, stability of optimal solutions with respect to parameters, the construction of optimization algorithms, their application to concrete problems. We consider that our work has three components :
Basic research which can be motivated by advanced applications, or by ongoing work among the international scientific community.
Development of optimization algorithms, aimed at solving large classes of emerging problems, rather than particular instances of applications.
Applications This means the collaboration with industrial partners (or coming from other research areas) on specific problems they are faced with.
Naturally, these three domains are not independent : each one is constantly nourrished by the other two. This year, our main results concerned
-- the theory of optimization : 2nd-order development of convex functions, optimization on the set of positivie definite matrices, generic classification of control problems ;
-- applications : optimal management of electrical power-plants.