Hierarchical stratification of Pareto sets

In smooth and convex multiobjective optimization problems the set of Pareto optima is diffeomorphic to an $m-1$ dimensional simplex, where $m$ is the number of objective functions. The vertices of the simplex are the optima of the individual functions and the $(k-1)$-dimensional facets are the Pareto optimal set of $k$ functions subproblems. Such a hierarchy of submanifolds is a geometrical object called stratification and the union of such manifolds, in this case the set of Pareto optima, is called a stratified set. We discuss how these geometrical structures generalize in the non convex cases, we survey the known results and deduce possible suggestions for the design of dedicated optimization strategies.

No reviews yet.

No comments yet.