Training Algorithm of BP Networks Based on Improved Ant Colony Algorithm

Abstract:

This study proposes a new method to search optimal parameters of BP networks based on improved ant colony algorithm. The algorithm proposed that each ant searches only around the best solution of the previous iteration with, which can reduce search space fast. is proposed for improving the solution performance to reach global optimum fairly quickly. Simulation results indicate that optimize parameters of BP networks with this method can not only overcome the limitations both the slow convergence and the local extreme values by basic BP algorithm, but also improve the learning ability and generalization ability.

• Periodical: Advanced Materials Research (Volumes 204-210)
• Pages: 310-313
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.204-210.310
• Online since: February 2011
• Authors: Yu Yan Ren, Ming Sun, Jie Bao, Hong Rui Wang
• Keywords: Ant Colony Algorithm (ACA), BP Network, Optimization, Parameter

References

[1] Li Zuoyong and Peng Lihong: submitted to Science in China Ser. F Information Science (2004).

[2] Li Zuoyong, Wang Jiayang and Guo Chun: submitted to Journal of Electronics (2008).

[3] Song Naihua and Xing Qinghua: submitted to Computer Engineering (2006).

[4] Wang Jing: submitted to Computer engineering and application (2006).

[5] M.D. Toksar: submitted to Applied Mathematics and computation (2006).

[6] M.D. Toksar: submitted to Journal of Computational and Applied Mathematics (2007).

[7] A.V. Donati, R. Montemanni, N. Casagrande, A.E. Rizzoli and L.M. Gambardella: submitted to European Journal of Operational Research (2008).

[8] C.B. Cheng and C.P. Mao: submitted to Mathematical and computer Modeling (2007).

[9] Ozgur Baskan, Soner Haldenbilen, Huseyin Celyan and Halim Ceylan: submitted to Applied Mathematics and Computation (2009).