Stochasticity and terminal stacking performance of import containers

The use of containers has decreased sea transportation costs and has enabled more cost effective transport. Moreover, containerisation has led to advantages in terms of handling processes at container terminals and protection of the goods inside. With an increasing amount of container terminals around the world, competition between terminals has risen. Therefore, it has become more and more important for terminal operators to conduct their work as efficient as possible.

Efficient terminal operations are hindered by stochasticity, which is defined as random and uncertain variation in a parameter value. Stochasticity can affect many aspects of terminal operations. It is for example present in arrival times of hinterland transport modes (trains, trucks and barges) and next mode of transport for import containers. Upon arrival of a container at the container yard, where containers are temporarily stored and await further transport, it is often unknown when the container will be picked up again and by which mode of transport. Stochasticity in arrival times of hinterland transport modes makes it harder for a terminal to organize the import yard in a way that the containers that will be retrieved first are on top of the stack.

Due to the lack of information on exact arrival times, requested containers are often stored underneath other containers, which will then need to be reshuffled to reach the requested container. This type of reshuffling is performed while a hinterland transport mode is waiting for the requested container, and therefore increases service times for the hinterland transport modes. Reshuffling performed during peak hours especially puts pressure on the terminal system, since the limited amount of available yard operating equipment is already scheduled for many tasks during peak hours.

The goal of this research is to quantify the relation between stochasticity and the performance of stacking operations of import containers. The main research question is:

What is the effect of stochasticity in import container information on stacking performance at a container terminal?

Indicators that represent stacking performance are based on a literature review. Based on this review the following performance indicators have been defined: the number of reshuffles, the truck service time and the maximum straddle carrier (SC) utilisation rate per hour.

The simulation model designed in this research is based on a SC operated multimodal container terminal. The simulation model is specified with data on terminal layout, retrieved volumes per hinterland transport mode over a period of 46 weeks, truck arrival distributions and average dwell time.

Different experiments have been run to test the effect of stochasticity on stacking performance. In the experiments two configurations have been specified: the implementation of housekeeping moves during non-peak hours (from 10:00 PM until 5:30 AM) and the non-housekeeping configuration. For these two configurations, results of experiments in which the level of stochasticity is varied are achieved.

The research has led to two main conclusions. The first conclusion is that stochasticity deteriorates stacking performance. The second conclusion is that implementing housekeeping moves improves the performance of stacking operations, even in scenarios with a high level of stochasticity. Implementing housekeeping moves in the scenario with low stochasticity could spare almost 640.000 reshuffles per year, almost 9.000 hours of truck service time per year and 8 minutes per SC in peak hours based on the total import volume of the terminal.

In the scenario with high level of stochasticity, implementing housekeeping moves could save more than 120.000 reshuffles per year, almost 900 hours of truck service time per year and 3 minutes per SC in peak hours. In the low stochasticity experiment 10 housekeeping SC shifts are needed each night to obtain the benefits and the total SC working hours increase with more than 60.000 hours per year for the entire import yard. In the high stochasticity experiment 8 housekeeping SC shifts are needed and the total SC working hours increase with almost 59.000 hours per year.

Read the full study by Nienke Mutters (TU Delft) here