등급기준 돌연변이 확률조절에 여왕벌진화의 융합을 통한 유전자알고리즘의 성능 향상

Performance Improvement of Genetic Algorithms through Fusion of Queen-bee Evolution into the Rank-based Control of Mutation Probability

  • 투고 : 2012.02.29
  • 심사 : 2012.06.21
  • 발행 : 2012.07.25

초록

본 논문에서는 기 개발된 등급기준 돌연변이 확률조절방법에 여왕벌진화방법을 융합하여 유전자알고리즘의 성능을 향상시키는 방법을 제안한다. 등급기준 돌연변이 확률조절 방법은 유전자알고리즘의 개체가 지역 최적해에 빠지는 것을 방지하고 지역 최적해에 빠졌을 경우 쉽게 빠져나올 수 있게 하는 방법으로 기존 알고리즘에 비하여 일정부분 성능향상을 보였다. 그러나 이 방법은 지역최적해가 많건 적건 간에 전역 최적해가 한 곳에 작은 영역에 있는 문제에서는 그다지 성능이 좋지 않았다. 우리는 그 이유가 이 방법이 전역 최적해로의 수렴성이 부족한 것으로 판단하고 수렴성을 강화시키기 위하여 여왕벌 진화방법을 융합한 알고리즘을 본 논문에서 제안한다. 여왕벌진화방법은 여왕벌의 생식을 모사한 방법으로 수렴성을 강화시킬 수 있는 방법이다. 제안한 방법의 성능을 측정하기위하여 4개의 함수최적화문제에 적용해본 결과 우리가 예상한대로 전역 최적해가 한 곳에 작은 영역에 몰려있는 문제에서는 상당한 성능향상이 일어나는 것을 관찰할 수 있었다. 그러나 전역 최적해가 넓은 영역에 걸쳐있는 문제에서는 성능향상이 거의 없었으며 전역 최적해가 여러 곳에 멀리 떨어져 있는 문제에서는 강한 수렴성으로 인하여 오히려 성능이 나빠지는 것을 볼 수 있었다. 이러한 실험결과로 보았을 때 본 논문에서 제안한 방법은 전역 최적해가 한 곳에 몰려있는 문제에서 매우 유용하게 사용될 수 있을 것으로 판단된다.

This paper proposes a fusion method of the queen-bee evolution into the rank-based control of mutation probability for improving the performances of genetic algorithms. The rank-based control of mutation probability which showed some performance improvements than the original method was a method that prevented individuals of genetic algorithms from falling into local optimum areas and also made it possible for the individuals to get out of the local optimum areas if they fell into there. This method, however, showed not good performances at the optimization problems that had a global optimum located in a small area regardless of the number of local optimum areas. We think that this is because the method is insufficient in the convergence into the global optimum, so propose a fusion method of the queen-bee evolution into this method in this paper. The queen-bee evolution inspired by reproduction process of queen-bee is a method that can strengthen the convergency of genetic algorithms. From the extensive experiments with four function optimization problems in order to measure the performances of proposed method we could find that the performances of proposed method was considerably good at the optimization problems whose global optimum is located in a small area as we expected. Our method, however, showed not good performances at the problems whose global optima were distributed in broad ranges and even showed bad performances at the problems whose global optima were located far away. These results indicate that our method can be effectively used at the problems whose global optimum is located in a small area.

키워드

참고문헌

  1. D. Goldberg, "Genetic Algorithms in Search, Optimization and Machine Learning," Addison- Wesley, 1989.
  2. M. Srinivas and L. M. Patnaik, "Genetic Algorithms: A Survey," IEEE Computer Magazine, pp. 17-26, June 1994.
  3. H. Szczerbicka and M. Becker, "Genetic Algorithms: A Tool for Modelling, Simulation, and Optimization of Complex Systems," Cybernetics and Systems: An International Journal, vol. 29, pp. 639-659, Aug. 1998. https://doi.org/10.1080/019697298125461
  4. R. Yang and I. Douglas, "Simple Genetic Algorithm with Local Tuning: Efficient Global Optimizing Technique," Journal of Optimization Theory and Applications, vol. 98, pp. 449-465, Aug. 1998. https://doi.org/10.1023/A:1022697719738
  5. J. Andre, P. Siarry, and T. Dognon, "An improvement of the standard genetic algorithm fighting premature convergence in continuous optimization," Advances in engineering software, vol. 32, no. 1, pp. 49-60, 2001. https://doi.org/10.1016/S0965-9978(00)00070-3
  6. C. Xudong, Q. Jingen, N. Guangzheng, Y. Shiyou, and Z. Mingliu, "An Improved Genetic Algorithm for Global Optimization of Electromagnetic Problems," IEEE Transactions on Magnetics, vol. 37, pp. 3579-3583, Sept. 2001. https://doi.org/10.1109/20.952666
  7. J. A. Vasconcelos, J. A. Ramirez, R. H. C. Takahashi, and R. R. Saldanha, "Improvements in Genetic Algorithms," IEEE Transactions on Magnetics, vol. 37, pp. 3414-3417, Sept. 2001. https://doi.org/10.1109/20.952626
  8. S. H. Jung, "Queen-bee evolution for genetic algorithms," Electronics Letters, vol. 39, no. 6, pp. 575-576, Mar. 2003. https://doi.org/10.1049/el:20030383
  9. E. Alba and B. Dorronsoro, "The exploration/ exploitation tradeoff in dynamic cellular genetic algorithms," IEEE Transactions on Evolutionary Computation, vol. 9, pp. 126-142, Apr. 2005. https://doi.org/10.1109/TEVC.2005.843751
  10. V. K. Koumousis and C. Katsaras, "A saw-tooth genetic algorithm combining the effects of variable population size and reinitialization to enhance performance," IEEE Transactions on Evolutionary Computation, vol. 10, pp. 19-28, Feb. 2006.
  11. A. E. Eiben, Z. Michalewicz, m. Schoenauer, and J. E. Smith "Parameter Control in Evolutionary Algorithms," Studies in Computational Intelligence, vol. 54, pp. 19-46, 2007.
  12. Silja Meyer-Nieberg and Hans-Georg Beyer, "Self-Adaptation in Evolutionary Algorithms," Studies in Computational Intelligence, vol. 54, pp. 47-75, 2007.
  13. S. H. Jung, "Rank-based Control of Mutation Probability for Genetic Algorithms," International Journal of Fuzzy Logic and Intelligent Systems, vol. 10, no. 2, pp. 146-151, May 2010. https://doi.org/10.5391/IJFIS.2010.10.2.146
  14. C. W. Ho, K. H. Lee, and K. S. Leung, "A Genetic Algorithm Based on Mutation and Crossover with Adaptive Probabilities," in Proceedings of the 1999 Congress on Evolutionary Computation, vol. 1, pp. 768-775, 1999.
  15. M. Srinivas and L. M. Patnaik, "Adaptive Probabilities of Crossover and Mutation in Genetic Algorithms," IEEE Transactions on Systems, Man and Cybernetics, vol. 24, no. 4, pp. 656-667, Apr. 1994. https://doi.org/10.1109/21.286385