Evaluate the Performance of Port Container Using an Hybrid Framework

Abstract

This work intends to integrate feedforward neural network (FNN) and data envelopment analysis (DEA) in a single framework to evaluate the performance of operations in the port container terminal. The proposed framework is based on three steps. In the first step, we identify the performance measures objectives and the indicators affecting our system. In the second step, a DEA-based oriented inputs model (DEA-CCR) is used to compute the efficiency scores of the system, based on the obtained scores, the data is divided into training and testing datasets. In the last step, an improved crow search algorithm (ICSA) is employed as a new method for training FNNs to determine the efficiency scores. In ICSA, the so-called Levy flights are used to enhance the convergence rate of CSA and prevent it from getting stuck in local optima. To demonstrate the efficacy of the proposed framework, it is utilized to evaluate the performance of two ports container terminal mainly: Tangier and Casablanca. The results are compared with a standard BBO, GA and PSO-based learning algorithm. The new trainer ICSA is also investigated and evaluated using four different classification datasets selected from the UCI machine learning repository and on three approximation functions datasets. The experimental results show that ICSA outperforms both BBO, GA and PSO for training FNNs in terms of converging speed and avoiding local minima.