The so-called “last mile” is the bane of delivery services. It is the final segment of a product’s journey to a consumer, the part of the trip that drops the parcel at a doorstep, at a pick-up locker or at a business address. It is the most expensive leg in the movement of goods.
It is also the subject of an international research collaboration involving Associate Professor of Decision Sciences Long He, whose expertise includes Smart City operations and supply chain management. In collaboration with Quanmeng Wang and Chung-Piaw Teo at the National University of Singapore Business School, He developed a mathematical model to measure whether the “last mile” could be optimized by delivering parcels to public storage lockers for pickup by customers.
The focus of the research was the Locker Alliance Network supported by the government of Singapore. But He said the results would be applicable to other large cities with locker systems, including Hong Kong, New York and D.C. The key is to set up locker networks that are “seen as a public good” and, therefore, accessible to all companies, much as Tesla car charging stations in some locations can now be used by non-Tesla vehicles.
“In many places, we are facing the challenge of last-mile delivery, the touchpoint where we see the UPS drivers, where we see the Amazon delivery trucks. It is the component that makes up 53 percent of the cost of logistics,” He said.
Asia is experimenting with models to improve last-mile efficiency. The Singapore government tried its hand at increasing efficiency through the creation of the Locker Alliance Net, a system of 1,000 lockers it set up across two towns in 2021.
“The network is a public system of lockers to which all companies and providers have access,” He said. “Any company has access to that locker. Any company can deliver to that locker. The study we did was in a town with 39 lockers in a residential area … with relatively cheaper housing and a [public transit] system. They tried to install the lockers in convenient locations.”
He said a public locker network offers three notable benefits. If reduces the number of failed deliveries. If a package is delivered to a home and no one is there to receive it, it may have to be redelivered at a later time, which adds costs to the delivery providers. Public lockers bypass this problem.
Public lockers also provide a way to consolidate deliveries, so a truck driver can leave everything in one location rather than driving from home to home. The third benefit is that it is interoperable—every company doesn’t need its own Amazon-style locker system. They can all share one network, which lowers costs.
In their study, He and his colleagues developed mathematical models to analyze whether the current design of the Singapore network, which includes both locker delivery and home delivery, really improved last-mile efficiency. He said the goal was to find out if “we are creating more costs or are we driving more miles than a system without lockers?”
Both customer behavior and the locations of the lockers were built into the analysis. The researchers looked at whether doing the home deliveries separate from the locker deliveries, or combining them, was more efficient. They also looked at routing for the deliveries.
What they found was that Singapore’s current locker system, with deliveries made separately to homes and lockers, increased the delivery distance by 30 percent. By moving the locker locations, however, the trip time could be cut by 2 percent.
“Our findings were striking,” He said.
He noted that 40 percent of consumers would need to use the redesigned locker system to optimize the results. However, under the existing system in which separate trips are made to homes and to lockers, 60 percent of consumers need to use the locker system.
“We looked at the problem to see if there was a better design. In the Singapore case, they needed to improve the locker-use adoption rate by consumers and they needed to improve the locker locations,” He said.
In the United States, he said, officials could look at whether Amazon, for example, would be willing to allow UPS or FedEx to use its lockers and whether the Amazon lockers are located to optimize delivery efficiency. A separate challenge is how to change consumer behavior so pickup at lockers seems more desirable. He said researchers in Asia are currently looking into what factors improve consumer use of lockers.
He’s work on social and public infrastructure efficiency also includes research on bike-sharing systems and more advanced AI techniques to predict customer preferences.