Using predictive analytics, the Port of Montreal this year will alert drayage drivers when it looks as if truck turn times will be long, and use data collected on terminal gate movements to lessen the impact of rail switching on trucks. Canada’s second-largest port and the city also will use data collected by radio-frequency identification and license plate reading technology to manage traffic signals, improving truck and passenger traffic fluidity.

Drayage drivers serving Montreal’s five container terminals and their dispatchers have been able to monitor truck turn times in real time since September 2016. Predictive analysis will build on the information collected and available via the Trucking PORTal application by enabling truckers to receive notifications if the models suggest the wait would exceed what the driver was willing to endure. Accurate forecasts of turn times would enable truckers to determine whether their time would be better spent doing a pickup or drop-off at another terminal.

“Drivers and dispatchers have a personalized account in which they can set alerts themselves with their own thresholds, depending on their individual tolerance margin,” said Daniel Olivier, the port’s director of business intelligence and innovation. “As for the predictive analytics, we are still testing various scenarios, but the time horizon for now is up to 48 hours. The further out you try to forecast, the quality of the forecast will be affected downward.”

A local technology startup specializing in artificial intelligence (AI) is helping the port with the C$1.3 million project ($1.05 million), which Transport Canada is funding and running jointly with the terminals. The port’s effort fits well with the city’s rising status as a hub for AI development and Canada’s broader effort to reduce greenhouse gases by reducing truck idling.

The need for better visibility to make smarter decisions, along with better coordinating of train switching and traffic flows on local roads, comes as North American port volume rises. And, although Montreal does not have the same intensity of import discharges as other major ports because depth limitations on the St. Lawrence Seaway prevent mega-ships from calling, the port is no different from others in its need to ensure speedy, reliable access for truckers on behalf of their shipper and consignee customers. 

Of the 2,500 trucks that enter the port daily, about 1,700 haul containers, and the share of cargo moved by truck versus rail has flipped from 45:55 to 55:45 in the last decade, according to Olivier. Volume is rising, too, with the port estimating it ended 2017 with 1.5 million TEU, up about 6 percent from 2016.

Despite the increasing pressure on terminal gates, the average truck turn time declined from 48 minutes in 2016 to 43 minutes in 2016, Olivier said. Truck turn times are measured by the time the vehicle enters the common entry portal, where the RFID tags are read, and then exits the facility, having dropped off a container and picking up another.

Montreal’s effort is another example of the potential to tap predictive analytics to move goods in and out of marine terminals more efficiently. In the US, chassis lessors are using data collection and predictive analytics to better match equipment to demand. And Advent Intermodal Solutions this month will roll out a tool at Long Beach’s International Transportation Service terminal that aims to give truckers and shippers a prediction of when import containers are available for pickup at marine terminals.

“The challenge of such a proactive approach is to consistently produce reasonably accurate forecasts to gain the trust of the drayage community,” said ManWo Ng, assistant professor of maritime and supply chain management at Old Dominion University. “At the end, while observed truck turn times are never wrong, predictions are most likely always wrong,”

More than 4,200 direct users visited the Trucking PORTal application in the first year following its September 2016 launch, according to Olivier. Roughly 80 percent of the logins were made via desktop devices, suggesting that dispatchers are making use of it, but there’s still room for more truckers to tap the app via their smartphones.

The potential for predictive analysis is exciting, though it’s important to note that even the best models can’t predict the future. “Although this is rather disappointing, one has to keep in mind that the benchmark should not be the unattainable truth, but rather the predictions that currently inform the decisions made in practice,” Ng said. “In other words, predictions have value as long as they are able to paint a clearer picture of the future than the image currently in place.”