A hybrid gradient method for vector optimization problems
摘要
In this work, we present and analyze a variant of the classical gradient method in Hilbert spaces for solving vector optimization problems where the objective function is Fréchet differentiable with L-Lipschitz continuous Jacobian map. The proposed algorithm utilizes a hybrid scheme involving a subproblem based on first-order information of the objective function and a projection step onto the intersection of two specific sets, which are constructed iteratively. This projection step ensures that the generated sequence exhibits strong convergence to a weakly efficient solution. Furthermore, a numerical experiment is provided to evaluate the practical performance of the Hybrid Vector Gradient Method (HVGM) based on the developed theory.