TY - JOUR
T1 - Predictive analytics for truck arrival time estimation
T2 - a field study at a European distribution center
AU - van der Spoel, Sjoerd
AU - Amrit, Chintan Amrit
AU - van Hillegersberg, Jos
N1 - Special issue: Using big data to make better decisions in the digital economy
PY - 2017
Y1 - 2017
N2 - Distribution centres (DCs) are the hubs connecting transport streams in the supply chain. The synchronisation of coming and going cargo at a DC requires reliable arrival times. To achieve this, a reliable method to predict arrival times is needed. A literature review was performed to find the factors that are reported to predict arrival time: congestion, weather, time of day and incidents. While travel time receives considerable attention, there is a gap in literature concerning arrival vs. travel/journey time prediction. None of the reviewed papers investigate arrival time: all the papers found investigate travel time. Arrival time is the consequence of travel time in combination with departure time, so though the travel time literature is applicable, the human factor involved in planning the time of departure can affect the arrival time (especially for truck drivers who have travelled the same route before). To validate the factors that influence arrival time, the authors conducted a detailed case study that includes a survey of 230 truckers, a data analysis and a data mining experiment, using real traffic and weather data. These show that although a ‘big data’ approach delivers valuable insights, the predictive power is not as high as expected; other factors, such as human or organisational factors, could influence arrival time, and it is concluded that such organisational factors should be considered in future predictive models.
AB - Distribution centres (DCs) are the hubs connecting transport streams in the supply chain. The synchronisation of coming and going cargo at a DC requires reliable arrival times. To achieve this, a reliable method to predict arrival times is needed. A literature review was performed to find the factors that are reported to predict arrival time: congestion, weather, time of day and incidents. While travel time receives considerable attention, there is a gap in literature concerning arrival vs. travel/journey time prediction. None of the reviewed papers investigate arrival time: all the papers found investigate travel time. Arrival time is the consequence of travel time in combination with departure time, so though the travel time literature is applicable, the human factor involved in planning the time of departure can affect the arrival time (especially for truck drivers who have travelled the same route before). To validate the factors that influence arrival time, the authors conducted a detailed case study that includes a survey of 230 truckers, a data analysis and a data mining experiment, using real traffic and weather data. These show that although a ‘big data’ approach delivers valuable insights, the predictive power is not as high as expected; other factors, such as human or organisational factors, could influence arrival time, and it is concluded that such organisational factors should be considered in future predictive models.
KW - IR-96577
KW - METIS-311007
U2 - 10.1080/00207543.2015.1064183
DO - 10.1080/00207543.2015.1064183
M3 - Article
VL - 55
SP - 5062
EP - 5078
JO - International journal of production research
JF - International journal of production research
SN - 0020-7543
IS - 17
ER -