Earticle

초록

영어

Due to the slow convergent speed of particle and easily get trapped in the local optima, a novel simple PSO algorithm with opposition-based learning average elite strategy is proposed. In this algorithm, a particle updating formula of the simplified swarm optimization (sPSO) algorithm is adopted. Moreover, the opposition-based learning elite strategy and Gaussian disturbance are exerted on the personal best particles and then replace personal best particle of sPSO with the average of elite opposite solutions with Gaussian disturbance of personal best particles. The adjustment of inertia weight is based on setting a threshold and then the inertia weight selects each mode adaptively according to its current state. A set of experimental results on benchmark functions demonstrate that the proposed PSO algorithm is an effective and efficient approach for optimization problems. Furthermore, the T-test analysis shows that the proposed algorithm is able to achieve better results.

목차

Abstract
 1. Introduction
 2. Basic Description of Simple PSO
 3. Design of OLAE-SPSO
  3.1. The Introduction of Elite Opposition-Based Learning Strategy
  3.2. The Addition of Improved Gauss Disturbance
  3.3. The Simplification of The Particle Updating Formula
  3.4. Decreasing Inertia Weight Based On Cosine Function
  3.5. Procedure of the OLAE-SPSO
  3.6. Runtime Complexity of the OLAE-SPSO
 4. Algorithm Simulation and Analysis
  4.1. Test Functions
  4.2. Experimental Parameters Settings
  4.3. Experimental Results and Discussion
  4.4. T-Test Analysis
 5. Conclusion
 Acknowledgements
 References

키워드

저자

• Bing AI [ College of Information Science and Engineering, Guilin University of Technology, Guilin 541004, China ]
• Ming-Gang DONG [ College of Information Science and Engineering, Guilin University of Technology, Guilin 541004, China ]
• Chuan-Xian JANG [ College of Information Science and Engineering, Guilin University of Technology, Guilin 541004, China ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of Hybrid Information Technology

• 간기

  격월간

• pISSN

  1738-9968

• 수록기간

  2008~2016

• 십진분류

  KDC 505 DDC 605