PSO-SVM Model Based Prediction and Analysis for the Formation of Navigation Channel Silt

Abstract:

Monitor and predict the change of navigation channel silt is important to ensure the safety of the channel while one of many difficulties is the deformation monitoring data is complicated and nonlinear, so its difficult to establish a deterministic model. Supporting vector machine could be widely used in the prediction of the formation of navigation channel silt because it has a good generalized ability, which could solve the problems like small sample, nonlinear, high-dimension. Because whether the algorithm could work or not based on the selection of the parameters, so a PSO-SVM based prediction model of the formation of the silt was established by using particle swarm optimization, which is a the fast overall optimization, and then was used to optimize the model parameter of the support vector machine. Study shows utilize this model in the silt formation in Huanghua harbor is plausible.

• Periodical: Applied Mechanics and Materials (Volumes 543-547)
• Pages: 4133-4136
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.543-547.4133
• Online since: March 2014
• Authors: Ping Wang*, Zhi Hong Qie, Fu Sheng Yang
• Keywords: Formation, Navigation, Prediction, PSO-SVM, Silt

References

[1] Longhai Zhu, Jianzheng Wu, Influence on channel siltation by sediment transport in Huanghua harbor silt dumping area. [J]，2010（11）：74-79.

[2] Hua Yang, ZhiQiang Hou, Analysis of rapid siltation of Huanghua harbor outer channel[J], Waterway and harbor, 2011（3）: 35-39.

[3] Chen Chen, Kai Zhu, A predictive investigation of the arch dam deformation based on PSO-SVM model[J], three gorges university journal，2013（2）：46-52.