In the last two decades, many descent methods for multiobjective optimization problems were proposed. In particular, the steepest descent and the Newton methods were studied for the unconstrained case. In both methods, the search directions are computed by solving convex subproblems, and the stepsizes are obtained by an Armijo-type line search. As a consequence, the objective function values decrease at each iteration of the algorithms. In this work, we consider nonmonotone line searches, i.e., we allow the increase of objective function values in some iterations. Two well-known types of nonmonotone line searches are considered here: the one that takes the maximum of recent function values, and the one that takes their average. We also propose a new nonmonotone technique specifically for multiobjective problems. Under reasonable assumptions, we prove that every accumulation point of the sequence produced by the nonmonotone version of the steepest descent and Newton methods is Pareto critical. Moreover, we present some numerical experiments, showing that the nonmonotone technique is also efficient in the multiobjective case.
Multiobjective optimization Steepest descent method Newton method Nonmonotone line search Pareto optimality
This is a preview of subscription content, log in to check access.
We would like to thank the anonymous referees for their suggestions, which improved the original version of the paper.
Das, I., Dennis, J.E.: Normal-boundary intersection: a new method for generating the Pareto surface in nonlinear multicriteria optimization problems. SIAM J. Optim. 8(3), 631–657 (1998)MathSciNetCrossRefzbMATHGoogle Scholar
Jin, Y., Olhofer, M., Sendhoff, B.: Dynamic weighted aggregation for evolutionary multi-objective optimization: Why does it work and how? In: GECCO’01 Proceedings of the 3rd Annual Conference on Genetic and Evolutionary Computation, pp. 1042–1049 (2001)Google Scholar
Kim, I.Y., de Weck, O.L.: Adaptive weighted-sum method for bi-objective optimization: Pareto front generation. Struct. Multidiscip. Optim. 29(2), 149–158 (2005)CrossRefGoogle Scholar
Laumanns, M., Thiele, L., Deb, K., Zitzler, E.: Combining convergence and diversity in evolutionary multiobjective optimization. Evolut. Comput. 10(3), 263–282 (2002)CrossRefGoogle Scholar
Mita, K.: Nonmonotone line search in multiobjective settings (in Japanese). Undergraduate research, Kyoto University(2017)Google Scholar
Mita, K., Fukuda, E.H., Yamashita, N.: On using nonmonotone line search techniques in steepest descent methods for multiobjective optimization (in Japanese). In: Proceedings of the 61st Annual Conference of the Institute of Systems, Control and Information Engineers (2017)Google Scholar
Ogata, Y., Saito, Y., Tanaka, T., Yamada, S.: Sublinear scalarization methods for sets with respect to set-relations. Linear Nonlinear Anal. 3(1), 121–132 (2017)MathSciNetzbMATHGoogle Scholar
Qu, S., Ji, Y., Jiang, J., Zhang, Q.: Nonmonotone gradient methods for vector optimization with a portfolio optimization application. Eur. J. Oper. Res. 263, 356–366 (2017)MathSciNetCrossRefzbMATHGoogle Scholar
Stadler, W., Dauer, J.: Multicriteria optimization in engineering: a tutorial and survey. In: Kamat, M.P. (ed.) Progress in Aeronautics and Astronautics: Structural Optimization: Status and Promise, vol. 150, pp. 209–249. American Institute of Aeronautics and Astronautics, Reston (1992)Google Scholar
Toint, Ph.L.: Test problems for partially separable optimization and results for the routine PSPMIN. Tech. Rep. 83/4, Department of Mathematics, University of Namur, Brussels (1983)