Artificial Bee Colony Optimized Recurrent Neural Network-Based Port Container Throughput Forecast

The Ports have become increasingly significant in the operation of international commercial activities as a result of the current economic globalisation. However, freight transportation planning for cargo movement through ports has become a big concern for the transportation sector over time and it is an essential duty for the country’s economic progress. Therefore, determining the feasibility of a strategic port development necessitates assessing the port throughput. However, container throughput data is complicated and it frequently has multiple seasons, which makes precise forecasting difficult. To enhance the accuracy of container throughput forecast of port, an enhanced Artificial Bee Colony based Recurrent Neural Network (ABC-RNN) prediction model was proposed. Twenty-four sets of data are created to create an RNN prediction model, and the network output is the container throughput statistics (2015–2019) of the VOC port in Tuticorin. The thresholds and weights of Recurrent Neural Network (RNN) are optimised and they are finally established. The ESN is used with ABC optimized RNN to further enhance the dynamic process of learning and biological system modelling. The study also compares forecasted foreign trade container volume estimates from the RNN, ABC and combination technique of ABC optimized RNN for the cities of Cochin and VOC from 2015 to 2019. The results demonstrates that, the combined prediction model of ABC optimized RNN effectively improves prediction precision and produce more precise outcomes.