摘要：AbstractThe seaport plays a critical role in the global supply chain management. Today, with the increasing growth of the global containerized trade, many challenges appears for the seaport, one of the most significant challenges is the berth allocation problem BAP, this fact impact all the seaport operations and benefits. The aim of this paper is to treat one the lead causes of BAP which is the berth congestion risk. Based on an ANN algorithm we tried to predict the congestion risk which is related directly with of the delay of vessels in the berth of the seaport. In this sense, we used a real combined dataset containing many features and we treat the correlation of each feature with the dependent variable, before applying six known models in order to compare the performance of the proposed ANN model. The result ensures that the proposed model surpasses the six traditional model on the tree chosen metrics of evaluation, F1 Score F1S = 86%, Recall RS = 88% and Accuracy = 96.5% which allow us to adopt this model for our case. In addition the developers can push the model even further, by using for example the natural inspired algorithms to enhance the accuracy.
