Separated Representations: A Powerful Strategy for Model Reduction

International audienceIn this paper we discuss about the features of a novel numerical method based on the use of separated representations for the variables of interest. This separated representation allows for an easy treatment of problems defined in a space of high number of dimensions, or in which some parameters are unknown or uncertain. As an example of such a problem we study the problem of cell signalling under a stochastic framework. The method also enables to solve problems in which one of the variables within is not known, by considering these unknowns as new model dimensions