A Barzilai-Borwein Descent Method for Multiobjective Optimization Problems

Jian Chen,Liping Tang,Xinmin Yang

DOI: https://doi.org/10.48550/arXiv.2204.08616

2022-04-19

Abstract:The steepest descent method proposed by Fliege et al. motivates the research on descent methods for multiobjective optimization, which has received increasing attention in recent years. However, empirical results show that the Armijo line search often gives a very small stepsize along the steepest direction, which decelerates the convergence seriously. This paper points out that the issue is mainly due to the imbalances among objective functions. To address this issue, we propose a Barzilai-Borwein descent method for multiobjective optimization (BBDMO) that dynamically tunes gradient magnitudes using Barzilai-Borwein's rule in direction-finding subproblem. With monotone and nonmonotone line search techniques, it is proved that accumulation points generated by BBDMO are Pareto critical points, respectively. Furthermore, theoretical results indicate the Armijo line search can achieve a better stepsize in BBDMO. Finally, comparative results of numerical experiments are reported to illustrate the efficiency of BBDMO and verify the theoretical results.

Optimization and Control

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the problem of overly small step - sizes in multi - objective optimization. Specifically, when dealing with multi - objective optimization problems, the traditional Steepest Descent Method for Multiobjective Optimization (SDMO), due to the imbalance between objective functions, often causes the Armijo line search to give very small step - sizes, which will seriously slow down the convergence speed of the algorithm. To solve this problem, the author proposes a Barzilai - Borwein Descent Method for Multiobjective Optimization (BBDMO). This method improves the direction - finding sub - problem by dynamically adjusting the magnitude of the gradient, so that the Armijo line search can obtain better step - sizes. The paper also proves that the accumulation points generated by BBDMO are Pareto critical points, and verifies the effectiveness of BBDMO through numerical experiments.