Capacity availability analysis for a seaport company

An important part of world trade is done by sea transportation. Therefore, ports have an important place in international trade. Capacity planning is very important in terms of the efficiency of the all enterprises and investment plans. Planning studies on capacity are based on demand estimation. The demand forecasting, which is a particularly important indicator for ports, provides an idea about determining the future capacity need. In this study, different methods are compared for demand forecasting problem of a seaport in Turkey. Three types of data were collected from the period of 2012-2017 for different operation types (general cargo, container, vehicle). When previous studies are examined, it is seen that machine learning methods are not used frequently for demand forecasting and studies on demand forecasting at container terminals are much more than studies conducted at marine terminals. Statistical forecasting methods and machine learning methods were applied to three data sets to determine the best estimation method and to estimate the handling volumes for the next two years after 2017. The comparison of the forecasting performances of statistical forecasting methods and machine learning methods have been analysed. According to chosen accuracy measures, Additive Holt Winter's was recognized as the best forecasting method for container and vehicle handling volumes, whereas ARIMA method ensured the best forecasting values for general cargo. For each data type, the prediction results of the best performing model were used to analyze whether the port has the capacity to handle the demand for the next two years after 2017 due to the quay construction work. According to the calculated capacity utilization rates, it will be seen that the quay to be demolished will affect the occupancy of the quay much and therefore plans should be made to increase the operational speed and increase storage areas until the pier construction is completed.