Read Full Paper

Using lockers for parcel delivery can significantly reduce congestion in city centers. This is among the major findings of research by HKUST’s Guodong Lyu and a colleague, focusing on a “smart nation” initiative recently proposed for implementation in Singapore. The researchers developed an innovative analytical method of determining the optimal network density and coverage of the proposed “Locker Alliance” network, based on a locker choice model. They show how useful insights can be gleaned from the model even without full transit route information on all citizens.

As e-commerce expands, businesses and residents increasingly rely on goods delivery. However, failed deliveries and wasted waiting hours are a major drag on the efficiency of this model. In cases of delivery failure, say the researchers, “companies have to either re-schedule another delivery or direct the customers to pick up the parcels at some designated place.” While as many as two thirds of residential customers still currently prefer home delivery, the shift to non-traditional pick-up locations may hold the key to overcoming these problems.

Coming on the heels of various private sector initiatives around the world, Singapore’s government has announced proposals for a “Locker Alliance” (LA) network to facilitate e-commerce. Based on automated infrastructure, and open to all parcel delivery firms, the network will provide an interoperable set of lockers, enabling customer pick-up from any location. By including residential zones, the network will “complement the existing infrastructure,” which, as the researchers note, mostly uses “proprietary lockers and collection points in commercial areas.”

Taking parcel delivery data from a Singapore company, the authors modeled the optimal design of the LA network in terms of two important questions regarding long-term viability. First, what should the network’s density and coverage be? To entice customers away from home delivery, lockers must be sufficiently accessible from their residence or workplace. Second, to what extent will lockers in residential blocks and rail stations divert demand from the central business district (CBD)? Reducing CBD congestion is one of the main expected benefits of this “smart nation” initiative.

Answering the first question, the researchers found that customers’ utility of locker usage—the factor driving their acceptance—was highly dependent on their distance from the lockers. “In particular,” write the authors, “the utility drops fast when the distance exceeds 250 meters.” Interestingly, 250 meters is exactly the maximum distance from residential blocks that the Singapore government aims to achieve for the LA network, showing that the proposal has been well thought out.

Addressing the second question, the researchers broke new ground by modeling the effect of the choice of pick-up location. When users can choose from 1,980 lockers across Singapore, residential customers will be particularly keen to use those in stations and shopping malls. However, such locations are prevalent in the CBD as well as residential areas. Hence, although the LA network will increase total locker usage, it could also stimulate parcel deliveries into the CBD. This unintended effect can be avoided by excluding lockers from that area.

Overall, the authors calculated that a well-designed network could reduce parcel delivery into the CBD of Singapore by at least 7.5%. To reap this benefit, they conclude that “planning agencies need to incentivize customers to use lockers, for example, by compensating the delivery fees.” The verified model developed in this study will aid policymaking and managerial decision-making on this and other such networks.