ORION: The algorithm proving that left isn't right


For UPS engineers, avoiding left turns isn't just a best practice for safety – it saves millions in dollars and gallons of fuel.

Long before there was an official policy at UPS, there was an unofficial understanding – no unnecessary left turns.

By the end of 2016, 55,000 ORION-optimized routes will have saved 10 million gallons of fuel annually and reduced 100,000 metric tons in CO2 emissions.

Left turns mean idling, which increases the time a route takes. Left turns mean going against traffic, which increases exposure to oncoming cars. Right turns are faster. Right turns save fuel.

Because most UPS managers have been UPS drivers, they have driven the routes and plotted on maps how to drive them with as many right-hand loops as possible. They knew right turns were the way to go, but that knowledge was in their heads.

"Before computers, engineering was about measurement and process," says Jack Levis, senior director of process management at UPS. "UPS has always believed in data, not intuition."

Eventually, UPS's technology caught up with experience. The result is ORION (or On-Road Integrated Optimization and Navigation). By optimizing delivery routes in regard to distance, fuel and time, ORION seeks to solve the Traveling Salesman Problem, which has stumped scientists for more than 200 years.

It seems simple: Start with a set of cities and find a route that goes through each city once, ending where it started. The problem gets interesting – and much more complex – when it tries to solve for the shortest among all possible routes. One solution is to map every round-trip route and find the shortest ones. The more cities, however, the more round-trips to map.

Ten cities mean more than 3.6 million possibilities. Sixteen cities? More than 20 trillion. One hundred cities? About 10156 circuits. If stops were cities, each UPS driver would visit at least 100 cities a day, and calculating the perfect route would be practically impossible – which is why ORION doesn't try to. 

ORION is a 1,000-page, algorithmic optimization. It's a technique for faster, more practical problem solving, providing a solution for an immediate situation, even if it is neither optimal nor perfect. ORION doesn't necessarily map the perfect route or even the best one. Rather, ORION gives UPS drivers workable routes, based on experience. It learns over time and speeds up the process. It gets smarter.

Ask consultants today, and they'll tell you there are three levels of analytics: descriptive, predicative and prescriptive. But in the early 1990s, when UPS moved from recording processes to analyzing data, the terms didn't exist.

"There were no handheld computers, so we built them," says Levis. Introduced in 1991, the Delivery Information Acquisition Device (DIAD) was the first iteration of the handheld data collector UPS drivers use to record and transmit delivery information across a mobile cellular network. By 1993 – and before smartphones – DIAD allowed drivers to record deliveries electronically, descriptively. But by the late 1990s, just tracking packages wasn't enough. UPS started examining the idea of connecting its physical network of hubs, spokes and drivers to a virtual network of the data from all those DIADs. The result: package flow technology, which uses smart labels to route packages to the right place on conveyer belts and, from there, to the best places on delivery trucks. By 2003, the ability to predict what would happen made the data about the package as important as the package itself. 

The third analytics phase – prescriptive – uses what happened and what will happen to guide decisions about what should happen. Enter ORION. In 2008, UPS piloted telematics – onboard data-gathering technology – to determine ways to improve efficiency. GPS tracking equipment and vehicle sensors, combined with drivers' handheld mobile devices, allowed UPS to capture data related to vehicle routes, idling times and safety protocols, such as seat belt use.

Implementing the solutions to gather the data called for an algorithm that could quickly solve complex routing problems. UPS tested the resulting ORION algorithm at various sites from 2003 to 2009; prototyped it at eight sites between 2010 and 2011; and deployed it to six beta sites in 2012. In a company with the size, scale and scope of UPS, even a small change can have huge implications. For UPS, eliminating one mile, per driver, per day over one year can save up to $50 million.

By the end of 2016, 55,000 ORION-optimized routes will have saved 10 million gallons of fuel annually, reduced 100,000 metric tons in CO2 emissions and saved an estimated $300 million to $400 million in cost avoidance. "ORION has been a game changer for UPS, impacting 55,000 drivers across 1,000 buildings in the United States when it is fully deployed," says Mark Wallace, senior vice president of global engineering and sustainability. "ORION has made us better at serving our customers – how they want, when they want, where they want."

ORION is the heart of the UPS My Choice® service, which provides online and mobile access to incoming UPS home deliveries, to set delivery preferences, reroute shipments and adjust delivery locations and dates as needed.

ORION is the result of a long-term operational technology investment and commitment by UPS, extending more than 10 years from the initial development of the algorithm to full deployment in nearly all 55,000 routes in the North American market by the end of 2016.

And ORION is just getting started. "Upcoming versions will use real-time data to anticipate bad weather, traffic or other route delays," says Levis.

In logistics, when shaving seconds can mean saving millions, optimization is everything – and not turning left is definitely the right thing to do.

Learn more about ORION here.


Reader Comments

Add Your Comment