Amazon has updated its algorithm to reduce delivery miles

According to a recent post on the Amazon Science website, Amazon expects the new delivery route algorithm to help it cut millions of miles of driving when it is deployed across the United States this year.

The Customer Order and Network Density Optimizer (Condor) algorithm evaluates the customer order and Network Density Optimizer (CRM) package before it leaves the fulfillment center to determine the most efficient shipping option. In January this year, the algorithm began a trial run in a few Amazon delivery stations, and then rolled out across the country. Now, Amazon plans to deploy it to other countries in the coming months. "We can enable operators to deliver more packages to more customers on time, while reducing the carbon emissions from driving miles and fuel," Andrea Quariza, Amazon's senior principal scientist, said in the article.

For Amazon, depending on the location of the item and other factors, it is possible to quickly decide how to fulfill a customer order. However, there may be a gap of 5-6 hours between the time an order is placed and when the fulfillment center starts processing the order, and if a nearby customer places an order later in the day, the order fulfillment plan may be updated. In the period leading up to the completion of the order, the customer order and the network density optimizer algorithm have the opportunity to re-evaluate the initial decision several times to see if a better route can be created.

The customer Order and network Density optimizer algorithms consider the entire geographic area served by the distribution station and determine how each delivery of the order is divided into shipping and procurement fulfillment centers. The breakthrough point of the program is that it is able to reduce the number of possible routing decisions per geographic block to less than 10, no matter how complex the situation.

In a test involving Amazon scientists and engineers, a control group of selected cities confirmed that customer orders and the network density Optimizer algorithm reduced required routing resources by 0.5%. Amazon chief scientist Rohit Marsch said: "This means that without Condor, the entire network would need 50,000 routes, and with the new algorithm, only 49,750 routes are needed."

Delivering the same number of packages with fewer routes and miles helps delivery companies save on "last mile" shipping costs, which is a big reason why UPS and fedex have launched initiatives in this area. UPS's recent pilot program includes holding orders for as long as the service agreement allows until they can be matched with another shipment at the same address. Through its "Web 2.0" initiative, fedex is integrating routes operated separately by express (that is, by air transport) companies and land carriers.