On the robustness of median sampling in noisy evolutionary optimization

Abstract

Evolutionary algorithms (EAs) are a sort of nature-inspired metaheuristics, which have wide applications in various practical optimization problems. In these problems, objective evaluations are usually inaccurate, because noise is almost inevitable in real world, and it is a crucial issue to weaken the negative effect caused by noise. Sampling is a popular strategy, which evaluates the objective a couple of times, and employs the mean of these evaluation results as an estimate of the objective value. In this work, we introduce a novel sampling method, median sampling, into EAs, and illustrate its properties and usefulness theoretically by solving OneMax, the problem of maximizing the number of 1s in a bit string. Instead of the mean, median sampling employs the median of the evaluation results as an estimate. Through rigorous theoretical analysis on OneMax under the commonly used onebit noise, we show that median sampling reduces the expected runtime exponentially. Next, through two special noise models, we show that when the 2-quantile of the noisy fitness increases with the true fitness, median sampling can be better than mean sampling; otherwise, it may fail and mean sampling can be better. The results may guide us to employ median sampling properly in practical applications.

References

  1. 1

    Bäck T. Evolutionary Algorithms in Theory and Practice: Evolution Strategies, Evolutionary Programming, Genetic Algorithms. Oxford: Oxford University Press, 1996

    Book  Google Scholar 

  2. 2

    Xu P, Liu X, Cao H F, et al. An efficient energy aware virtual network migration based on genetic algorithm. Front Comput Sci, 2019, 13: 440–442

    Article  Google Scholar 

  3. 3

    Yuan Q, Tang H B, You W, et al. Virtual network function scheduling via multilayer encoding genetic algorithm with distributed bandwidth allocation. Sci China Inf Sci, 2018, 61: 092107

    Article  Google Scholar 

  4. 4

    Jin Y C, Branke J. Evolutionary optimization in uncertain environments — a survey. IEEE Trans Evol Comput, 2005, 9: 303–317

    Article  Google Scholar 

  5. 5

    Aizawa A N, Wah B W. Scheduling of genetic algorithms in a noisy environment. Evolary Comput, 1994, 2: 97–122

    Article  Google Scholar 

  6. 6

    Stagge P. Averaging efficiently in the presence of noise. In: Proceedings of the 5th International Conference on Parallel Problem Solving from Nature, Amsterdam, 1998. 188–197

  7. 7

    Branke J, Schmidt C. Selection in the presence of noise. In: Proceedings of the 5th ACM Conference on Genetic and Evolutionary Computation, 2003. 766–777

  8. 8

    Branke J, Schmidt C. Sequential sampling in noisy environments. In: Proceedings of the 8th International Conference on Parallel Problem Solving from Nature, 2004. 202–211

  9. 9

    Auger A, Doerr B. Theory of Randomized Search Heuristics: Foundations and Recent Developments. Singapore: World Scientific, 2011

    Book  Google Scholar 

  10. 10

    Neumann F, Witt C. Bioinspired Computation in Combinatorial Optimization: Algorithms and Their Computational Complexity. Berlin: Springer, 2010

    Book  Google Scholar 

  11. 11

    Zhang Y S, Huang H, Wu H Y, et al. Theoretical analysis of the convergence property of a basic pigeon-inspired optimizer in a continuous search space. Sci China Inf Sci, 2019, 62: 070207

    MathSciNet  Article  Google Scholar 

  12. 12

    Hwang H-K, Witt C. Sharp bounds on the runtime of the (1+1) EA via drift analysis and analytic combinatorial tools. In: Proceedings of the 15th International Workshop on Foundations of Genetic Algorithms, 2019

  13. 13

    Huang H, Su J P, Zhang Y S, et al. An experimental method to estimate running time of evolutionary algorithms for continuous optimization. IEEE Trans Evol Comput, 2020, 24: 275–289

    Article  Google Scholar 

  14. 14

    Zhang Y A, Qin X F, Ma Q L, et al. Markov chain analysis of evolutionary algorithms on OneMax function — from coupon collector’s problem to (1+1) EA. Theory Comput Sci, 2020, 820: 26–44

    MathSciNet  Article  Google Scholar 

  15. 15

    Bian C, Qian C, Tang K. Towards a running time analysis of the (1+1)-EA for OneMax and LeadingOnes under general bit-wise noise. In: Proceedings of the 15th International Conference on Parallel Problem Solving from Nature, 2018. 165–177

  16. 16

    Dang-Nhu R, Dardinier T, Doerr B, et al. A new analysis method for evolutionary optimization of dynamic and noisy objective functions. In: Proceedings of the 20th ACM Conference on Genetic and Evolutionary Computation, 2018. 1467–1474

  17. 17

    Droste S. Analysis of the (1+1) EA for a noisy OneMax. In: Proceedings of the 6th ACM Conference on Genetic and Evolutionary Computation, 2004. 1088–1099

  18. 18

    Gießen C, Kötzing T. Robustness of populations in stochastic environments. Algorithmica, 2016, 75: 462–489

    MathSciNet  Article  Google Scholar 

  19. 19

    Qian C, Yu Y, Zhou Z H. Analyzing evolutionary optimization in noisy environments. Evolary Comput, 2018, 26: 1–41

    Article  Google Scholar 

  20. 20

    Sudholt D. On the robustness of evolutionary algorithms to noise: refined results and an example where noise helps. In: Proceedings of the 20th ACM Conference on Genetic and Evolutionary Computation, 2018. 1523–1530

  21. 21

    Qian C, Shi J-C, Yu Y, et al. Subset selection under noise. In: Proceedings of the 31st International Conference on Neural Information Processing Systems, 2017. 3563–3573

  22. 22

    Qian C. Distributed pareto optimization for large-scale noisy subset selection. IEEE Trans Evol Comput, 2020, 24: 694–707

    Article  Google Scholar 

  23. 23

    Dang D-C, Lehre P K. Efficient optimisation of noisy fitness functions with population-based evolutionary algorithms. In: Proceedings of the 13th International Workshop on Foundations of Genetic Algorithms, 2015. 62–68

  24. 24

    Prugel-Bennett A, Rowe J, Shapiro J. Run-time analysis of population-based evolutionary algorithm in noisy environments. In: Proceedings of the 13th International Workshop on Foundations of Genetic Algorithms, 2015. 69–75

  25. 25

    Qian C, Bian C, Yu Y, et al. Analysis of noisy evolutionary optimization when sampling fails. In: Proceedings of the 20th ACM Conference on Genetic and Evolutionary Computation, 2018. 1507–1514

  26. 26

    Qian C, Yu Y, Tang K, et al. On the effectiveness of sampling for evolutionary optimization in noisy environments. Evolary Comput, 2018, 26: 237–267

    Article  Google Scholar 

  27. 27

    Qian C, Bian C, Jiang W, et al. Running time analysis of the (1+1)-EA for OneMax and leadingones under bit-wise noise. Algorithmica, 2019, 81: 749–795

    MathSciNet  Article  Google Scholar 

  28. 28

    Friedrich T, Kotzing T, Krejca M S, et al. The compact genetic algorithm is efficient under extreme gaussian noise. IEEE Trans Evol Comput, 2017, 21: 477–490

    Google Scholar 

  29. 29

    Doerr B, Hota A, Kötzing T. Ants easily solve stochastic shortest path problems. In: Proceedings of the 14th ACM Conference on Genetic and Evolutionary Computation, 2012. 17–24

  30. 30

    Feldmann M, Kötzing T. Optimizing expected path lengths with ant colony optimization using fitness proportional update. In: Proceedings of the 12th International Workshop on Foundations of Genetic Algorithms, 2013. 65–74

  31. 31

    Friedrich T, Kötzing T, Krejca M S, et al. Robustness of ant colony optimization to noise. Evolary Comput, 2016, 24: 237–254

    Article  Google Scholar 

  32. 32

    Sudholt D, Thyssen C. A simple ant colony optimizer for stochastic shortest path problems. Algorithmica, 2012, 64: 643–672

    MathSciNet  Article  Google Scholar 

  33. 33

    Akimoto Y, Astete-Morales S, Teytaud O. Analysis of runtime of optimization algorithms for noisy functions over discrete codomains. Theory Comput Sci, 2015, 605: 42–50

    MathSciNet  Article  Google Scholar 

  34. 34

    Huber P, Ronchetti M. Robust Statistics. Hoboken: John Wiley & Sons, 2009

    Book  Google Scholar 

  35. 35

    Leys C, Ley C, Klein O, et al. Detecting outliers: do not use standard deviation around the mean, use absolute deviation around the median. J Exp Social Psychol, 2013, 49: 764–766

    Article  Google Scholar 

  36. 36

    DeNavas-Walt C, Proctor B, Smith J. Income, poverty, and health insurance coverage in the United States: 2011. U.S. Census Bureau, 2012

  37. 37

    Doerr B, Sutton A. When resampling to cope with noise, use median, not mean. In: Proceedings of the 21st ACM Conference on Genetic and Evolutionary Computation, 2019. 242–248

  38. 38

    Droste S, Jansen T, Wegener I. On the analysis of the (1+1) evolutionary algorithm. Theory Comput Sci, 2002, 276: 51–81

    MathSciNet  Article  Google Scholar 

  39. 39

    He J, Yao X. Drift analysis and average time complexity of evolutionary algorithms. Artif Intell, 2001, 127: 57–85

    MathSciNet  Article  Google Scholar 

Download references

Acknowledgements

This work was supported by National Key Research and Development Program of China (Grant No. 2017YFB1003102), National Natural Science Foundation of China (Grant Nos. 62022039, 61672478, 61876077), and MOE University Scientific-Technological Innovation Plan Program. The authors would like to thank the anonymous reviewers for their helpful comments and suggestions to this work.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Chao Qian.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Bian, C., Qian, C., Yu, Y. et al. On the robustness of median sampling in noisy evolutionary optimization. Sci. China Inf. Sci. 64, 150103 (2021). https://doi.org/10.1007/s11432-020-3114-y

Download citation

Keywords

  • evolutionary algorithms
  • noisy optimization
  • median sampling
  • computational complexity
  • runtime analysis