Solving the Maximum Exploitable Potential of Channel Bearing Capacity by Improved GA-BP Algorithm

Abstract

The existing researches mainly focus on the stable navigation depth estimation or scheme comparison, lacking of methods for predicting the capacity of waterways, methods. The prediction has problems such as the inability to consider various uncertain factors that affect the capacity of the channel, excessive subjective factors and the cumbersome prediction process. An improved GA-BP algorithm is proposed to predict the maximum exploitable potential of the channel capacity. On the basis of the traditional GA-BP algorithm, two edge successive correction algorithms are introduced to initially optimize the predictive network model weights and thresholds. Introducing improved cross mutation operator to improve population diversity and effectively expand the range of solution search. Finally, the test samples verify the rationality of the optimized network model. Compare with GA-BP algorithm and BP neural network, the results reflect that the proposed GA-BP algorithm can comprehensively reflect various uncertainties of the channel capacity and reduce the influence of subjective factors on the prediction results. The algorithm can effectively predict the maximum development potential of the channel capacity, and is more effective than traditional prediction algorithms.