The New Adaptive ETLBO Algorithms with K-Armed Bandit Model
TLBO is a novel efficient swarm intelligent algorithm. In this paper, we first analyze TLBO and ETLBO algorithms in detail. Aiming at the disadvantage of ETLBO that it has to adjust the number of elite according to the different problems, we propose an improved adaptive ETLBO algorithm AETLBO-KAB that is based on K-armed bandit model. Experiments are carried out on six popular continuous non-linear test functions, and the results show that AETLBO-KAB algorithm is effective and brings dramatic improvement compared with TLBO and ETLBO. Furthermore, a new perturbation strategy—discussion group strategy is proposed. And the experimental results indicate that the efficiency of AETLBO-KAB with discussion group algorithm exceeds AETLBO-KAB algorithm.
KeywordsConstraint optimization ETLBO K-armed bandit Adaptive
This work is supported by the National Natural Science Foundation of China (61170314,61373052), the Project of Jilin Provincial Science and Technology Development(20170414004GH).
- 1.Kopec, D., Lucci, S.: Artificial Intelligence in the 21st Century. Mercury Learning and Information, Dulles (2012)Google Scholar
- 7.Kennedy, J., Eberhart, R.C.: Particle swarm optimization. In: Proceedings of IEEE International Conference on Neural Networks, pp. 1942–1948. IEEE Press, Piscataway (1995)Google Scholar
- 8.Dorigo, M., Maniezzo, V., Colorni, A., et al.: Positive feedback as a search strategy (1991)Google Scholar
- 15.Zhou, Z.: Machine Learning. Tsinghua University Press, Beijing (2016)Google Scholar
- 17.Tang, K., Yáo, X., Suganthan, P.N., et al.: Benchmark functions for the CEC’2008 special session and competition on large scale global optimization, pp. 153–177. Nature Inspired Computation and Applications Laboratory, USTC, China (2007)Google Scholar