United Parcel Service rolls out self-learning system to speed shipping

United Parcel service (UPS) has begun rolling out software that builds upon drivers' interactions with its tracking system to map out the fastest and most efficient routes.

UPS began implementing the new system, called on-road integrated optimization and navigation (ORION), in October. The rollout of ORION will reach 10,000 delivery routes by the end of this year, and it is hoped will significantly reduce the miles driven by UPS drivers.  U.S. deployment to nearly all 55,000 routes is planned to be completed in 2017 and global deployments are planned for the future.

The company hopes to save more than 1.5 million gallons of fuel -- a reduction of just one mile each day per driver over the course of a year saves up to $50 million annually, UPS estimates.

The unique feature in the software is that it is self-learning, meaning that as on-the-ground situations change along routes (such as road work or change in customer preference), the software will take the new data and adapt its recommendations. Writing in Wired, Marcus Wohlsen points out that ORION moves beyond simple arithmetic calculations to employ heuristics, "the field of math and computer science devoted to finding answers that are good enough, and that get better based on past experience."

Simple math wouldn't work too well when so much big data is involved. UPS says that ORION consists of more than 250 million address data points. The software combines customer's shipping requirements with customized map data the company has compiled to provide UPS drivers with optimized routing instructions that meet service-level requirements. "ORION is constantly evaluating routing options up to the moment a driver leaves the UPS facility. At full deployment it will be running tens of thousands of route optimizations per minute."

Speaking to Wired, Jack Levis, UPS director of process management, is quoted as saying "UPS doesn’t discount the value of driver wisdom accumulated during years on a route. The best system, he says, is one that relies on both human and algorithmic intelligence, not just one or the other."

UPS says ORION is the result of a long-term operational technology commitment, more than a decade in the making. To gather the necessary data, UPS operations research scientists began piloting telematics technologies with the installation of advanced GPS tracking equipment and vehicle sensors in 2008. The integration of these technologies allows UPS to capture data related to vehicle routes and performance and driver safety. Driver handheld mobile devices and telematics technologies combined with custom mapping data and ORION algorithms provide more efficient routes for UPS drivers.

UPS says it already prototyped ORION technology at 11 different sites between 2008 and 2011 and engaged senior UPS drivers to "beat the computer." This challenge helped to identify business rules to bolster the algorithm.

Ultimately, as in all things related to Big Data, the accuracy and trustworthiness of the data is paramount. That's why UPS wants the software to keep learning, building upon drivers' experiences with route obstacles or changes in customer preferences (maybe they want deliveries sent to the back door instead of the front, for example). As noted in a report by The Washington Post's Mohana Ravindranath, UPS will keep tweaking the algorithm, and still enable drivers to override the system.

This post was originally published on Smartplanet.com