A subgradient method for multiobjective optimization on Riemannian manifolds (Q382921): Difference between revisions

The authors propose a subgradient-type method for solving a multiobjective optimization problem whose objective function, \(F=(f_1,\dots,f_m) : M \to \mathbb{R}^m\), is convex on an \(n\)-dimensional Riemannian manifold \(M\) (i.e., componentwise convex along geodesic segments joining any points of \(M\)). If this iterative method stops at some iteration, then it generates a Pareto critical point, which is actually a weak Pareto optimal point. Otherwise, it generates a sequence of points, which converges to a Pareto optimal point under appropriate assumptions.