practice with the restriction of the dwell time and terminal access. He also proposed that the extension of gate hours on a marine terminal can reduce the container dwell time.

Huang (2008) has proven that increased container dwell times lead to more unproductive moves that result in a decrease in the terminal’s efficiency in a very costly manner. Some of the main factors influencing DT that were identified in the literature are: 1) the location of the terminal; 2) the efficiency of terminal operations; 3) the implemented port policies such as monetary penalties for delayed shipments or extended gate hours; 4) customs;

5) the freight forwarder or the shipping company; 6) the available hinterland connections; 7) the mode of transport used; 8) the cargo being transferred; and 8) the business relationships developed between the involved parties (Moini et al. (2008); Rodrigue et al. (2008)).

Moini et al (2008) applied genetic algorithms to evaluate the main factors affecting the dwell time of containers and measured their impact on the terminal productivity. Furthermore, she highlighted the importance of acquiring data on the landside recipients and on the type of the transported goods. This information is expected to enhance the predictability of the proposed models. In addition, Moini (2010) established a relationship between truck gate activities and drayage operations at a marine container terminal using both analytical and simulation approaches. By applying data mining techniques she identified the importance of the abovementioned determinants on the DT. Towards the same direction Kourounioti et al (2015) proposed the development of a methodological framework that combines aggregate and disaggregate models aiming to predict the dwell time of containers in a marine terminal. For this purpose regression models were developed that showed the influence of a container’s consignee and commodity on the DT.

In addition, if the exact day a container was to be discharged from the terminal was known in advance, operators would be able to organize the yard appropriately so as to be able to retrieve containers with higher pick-up probabilities more easily, reduce rehandling moves and get full advantage of the available capacity. The importance of this information has also been highlighted in Zhao and Goodchild (2010) who developed a simulation model to evaluate how the use of information affects the efficiency of a marine terminal. The results illustrated that when the day of a truck’s arrival was known in advance there was a substantial decrease of non-productive moves.

In order to deal with the lack of informational flow several container port terminals have implemented Truck Appointment Systems (TAS). TAS is mainly a system which books a slot for a certain number (restricted by each terminal’s capacity) of binding transactions during a predefined time period (usually one hour). One of the first TAS was implemented on the marine terminals of Los Angeles and Long Beach in order to deal with the issues of traffic congestion and air pollution (Giuliano and O’ Brien, 2007).

The summarized reflection is that limited research exists on quantifying the factors influencing DT. It cannot be easily contradicted that knowing when a container will be picked-up from a seaport terminal is expected to assist significantly decision making in a tactical and operational level when designing terminal policies as well as in a strategic level when taking investment decisions.

3. Methodological framework

In this section, the methodological framework applied in this research is presented. Based on the literature review findings (Moini et al., (2008); Moini, (2010); Rodrigue, (2008)) we assume that the factors affecting DT are pided in three distinctive categories as follows:

1. Information related to the container such as

● Container’s size (20ft, 40ft)

● Container’s status (full, empty)

● Container’s type (reefer, general cargo, hazardous)

● Commodity