Reinforcement learning aided parameter control in multi-objective evolutionary algorithm based on decomposition

被引：12

作者：

Ning W. ^{[1
]}

Guo B. ^{[1
]}

Guo X. ^{[1
]}

Li C. ^{[1
]}

Yan Y. ^{[1
]}

机构：

[1] School of Aerospace Science and Technology, Xidian University, Xi’an

来源：

Progress in Artificial Intelligence | 2018年 / 7卷 / 4期

基金：

中国国家自然科学基金;

关键词：

Decomposition; Multi-objective optimization; Parameter control; Reinforcement learning;

D O I：

10.1007/s13748-018-0155-7

中图分类号：

学科分类号：

摘要：

Multi-objective evolutionary algorithm based on decomposition (MOEA/D) has been successfully applied in solving multi-objective optimization problems. However, the performance of MOEA/D could be severely influenced by its parameter settings. In this paper, we introduce reinforcement learning into MOEA/D as a generic parameter controller. The resulting algorithm, reinforcement learning enhanced MOEA/D (RL-MOEA/D), is used to adaptively control the neighborhood size T and the differential evolutionary operators used in MOEA/D. RL-MOEA/D is first compared with MOEA/D with a random parameter control mechanism and MOEA/Ds with some fixed parameter settings on ten widely used multi-objective test instances. Then, RL-MOEA/D is compared with FRRMAB to show the effectiveness of the proposed algorithm. The experimental results indicate that RL-MOEA/D is very competitive. Finally, the characteristics of RL-MOEA/D are studied. © 2018, Springer-Verlag GmbH Germany, part of Springer Nature.

引用

页码：385 / 398

页数：13

共 23 条

[1]

Bosman P.A.N., Thierens D., The balance between proximity and diversity in multiobjective evolutionary algorithms, IEEE Trans. Evol. Comput., 7, 2, pp. 174-188, (2003)

[2]

Consoli P.A., Mei Y., Minku L.L., Yao X., Dynamic selection of evolutionary operators based on online learning and fitness landscape analysis, Soft. Comput., 20, 10, pp. 3889-3914, (2016)

[3]

Eiben A.E., Horvath M., Kowalczyk W., Schut M.C., Reinforcement learning for online control of evolutionary algorithms, International Workshop on Engineering Self-Organising Applications, pp. 151-160, (2006)

[4]

Ginley B.M., Maher J., O'Riordan C., Morgan F., Maintaining healthy population diversity using adaptive crossover, mutation, and selection, IEEE Trans. Evol. Comput., 15, 5, pp. 692-714, (2011)

[5]

Goncalves R.A., Almeida C.P., Pozo A., Upper Confidence Bound (UCB) Algorithms for Adaptive Operator Selection in MOEA/D, Lecture Notes in Computer Science, pp. 411-425, (2015)

[6]

Karafotias G., Eiben A.E., Hoogendoorn M., Generic parameter control with reinforcement learning, Conference on Genetic and Evolutionary Computation, pp. 1319-1326, (2014)

[7]

Karafotias G., Hoogendoorn M., Eiben A., Parameter control in evolutionary algorithms: trends and challenges, IEEE Trans. Evol. Comput., 19, 2, pp. 167-187, (2015)

[8]

Karafotias G., Hoogendoorn M., Eiben A.E., Evaluating Reward Definitions for Parameter Control, Applications of Evolutionary Computation, pp. 667-680, (2015)

[9]

Li K., Fialho A., Kwong S., Zhang Q., Adaptive operator selection with bandits for a multiobjective evolutionary algorithm based on decomposition, IEEE Trans. Evol. Comput., 18, 1, pp. 114-130, (2014)

[10]

Lin Q., Liu Z., Yan Q., Du Z., Coello C.A.C., Liang Z., Wang W., Chen J., Adaptive composite operator selection and parameter control for multiobjective evolutionary algorithm, Inf. Sci., 339, pp. 332-352, (2016)

← 1 2 3 →