A comparison of quality measures for model selection in surrogate-assisted evolutionary algorithm
- 192 Downloads
Choosing a proper approximation model should be the first and the most fundamental problem to be solved when dealing with surrogate-assisted evolutionary algorithms. Till now, most of the model selection methods emphasize on obtaining the best surrogate model basing on model accuracy assessments. As the population ranking is of the most important part in evolutionary optimization, the target function of surrogate model should focus on the right ranking of candidate solutions. Therefore, in this paper, we make a comparison study on several model quality measures which basically dedicated to measuring the capability of surrogate model in selecting and ranking the candidate solutions. In order to investigate the compatibility between accuracy assessments and ranking correlation methods, four algorithms with different model selection strategies based on different quality measures are designed and comparative study is made by contrasting them to three specific surrogate-assisted evolutionary algorithms as well as the standard particle swarm optimization. Simulation results on ten commonly used benchmark problems and one engineering case demonstrate the efficacy of the designed model selection strategies and meanwhile provide further insight into the three model quality measures studied in this paper in model selection.
KeywordsSurrogate Model selection Bootstrap sampling Quality measure Particle swarm optimization
This work is supported by National Natural Science Foundation of China (Grant Nos. 61472269 and 61403272) and the State Key Laboratory of Synthetical Automation for Process Industries, Northeastern University, China, as well as the Fund Program for the Scientific Activities of Selected Returned Overseas Professionals in Shanxi Province.
Compliance with ethical standards
Conflict of interest
Author Haibo Yu declares that he has no conflict of interest. Author Ying Tan declares that she has no conflict of interest. Author Chaoli Sun declares that she has no conflict of interest. Author Jianchao Zeng declares that he has no conflict of interest.
This article does not contain any studies with human participants or animals performed by any of the authors.
- Clerc M (1999) The swarm and the queen: towards a deterministic and adaptive particle swarm optimization. In: Proceedings of the 1999 congress on evolutionary computation-CEC99 (Cat. No. 99TH8406), vol 1953, pp 1–1957. https://doi.org/10.1109/cec.1999.785513
- Díaz-Manríquez A, Toscano-Pulido G, Gómez-Flores W (2011) On the selection of surrogate models in evolutionary optimization algorithms. In: 2011 IEEE congress of evolutionary computation (CEC), pp 2155–2162. https://doi.org/10.1109/cec.2011.5949881
- Ingu T, Takagi H (1999) Accelerating a GA convergence by fitting a single-peak function. In: Fuzzy systems conference proceedings, FUZZ-IEEE ‘99. 1999 IEEE International, vol 1413, pp 1415–1420. https://doi.org/10.1109/fuzzy.1999.790111
- Jin Y, Michael H (2003) Quality measures for approximate models in evolutionary computation. In: GECCO, pp 170–173Google Scholar
- Lendasse A, Wertz V, Verleysen M (2003) Model selection with cross-validations and bootstraps: application to time series prediction with RBFN models. In: Artificial neural networks and neural information processing–ICANN/ICONIP 2003, pp 573–580. Springer, BerlinGoogle Scholar
- Liang K-H, Yao X, Newton C (2000) Evolutionary search of approximated n-dimensional landscapes. Int J Knowl Based Intell Eng Syst 4:172–183Google Scholar
- Lim D, Ong Y-S, Jin Y, Sendhoff B (2007) A study on metamodeling techniques, ensembles, and multi-surrogates in evolutionary computation. Paper presented at the Proceedings of the 9th annual conference on Genetic and evolutionary computation, London, EnglandGoogle Scholar
- Lim D, Ong Y-S, Jin Y, Sendhoff B (2008) Evolutionary optimization with dynamic fidelity computational models. In: Huang D-S, Wunsch DC, Levine DS, Jo K-H (eds) Advanced intelligent computing theories and applications. With Aspects of artificial intelligence: 4th international conference on intelligent computing, ICIC 2008 Shanghai, China, September 15-18, 2008. Springer, Berlin, pp 235–242 https://doi.org/10.1007/978-3-540-85984-0_29
- Lophaven SN, Nielsen HB, Søndergaard J (2002) DACE-A Matlab Kriging toolbox, version 2.0Google Scholar
- Nakayama H, Arakawa M, Sasaki R (2001) A computational intelligence approach to optimization with unknown objective functions. In: Artificial neural networks—ICANN 2001. Springer, Berlin, pp 73–80Google Scholar
- Powell M (2001) Radial basis function methods for interpolation to functions of many variables. In: HERCMA. Citeseer, pp 2–24Google Scholar
- Shi L, Rasheed K (2008) ASAGA: an adaptive surrogate-assisted genetic algorithm. Paper presented at the Proceedings of the 10th annual conference on Genetic and evolutionary computation, Atlanta, GA, USAGoogle Scholar
- Shi L, Rasheed K (2010) A survey of fitness approximation methods applied in evolutionary algorithms. In: Tenne Y, Goh C-K (eds) Computational intelligence in expensive optimization problems. Springer, Berlin, pp 3–28. https://doi.org/10.1007/978-3-642-10701-6_1
- Yang D, Flockton SJ (1995) Evolutionary algorithms with a coarse-to-fine function smoothing. In: IEEE international conference on evolutionary computation. IEEE, pp 657–662Google Scholar
- Yu H, Tan Y, Sun C, Zeng J, Jin Y (2016) An adaptive model selection strategy for surrogate-assisted particle swarm optimization algorithm. In: 2016 IEEE symposium series on computational intelligence (SSCI), pp 1–8 https://doi.org/10.1109/ssci.2016.7850208