Regularization of Quasi -Newton Methods for Composite Multi-objective Problems

In this paper, we propose an efficient regular-ized quasi-Newton method for solving the unconstrained composite multiobjective optimization problem (UCMOP) where the objective functions are neither smooth nor convex. Under some suitable conditions, we establish that every accumulation point of the sequence produced by the proposed algorithm is a Pareto stationary point of the (UCMOP). We have conducted numerical experiments that show the validity of the proposed method.