Last Mile Innovation: The Case of the Locker Alliance Network

51 Pages Posted: 25 Oct 2019

See all articles by Guodong Lyu

Guodong Lyu

National University of Singapore (NUS) - NUS Business School

Chung-Piaw Teo

NUS Business School - Department of Decision Sciences

Date Written: October 17, 2019


Problem Definition: Under the smart nation initiative, the Singapore government has recently proposed the concept of "Locker Alliance'' (LA), an interoperable network of lockers in residential areas and train stations, to improve the efficiency of last mile parcel delivery operations. This is to complement the existing locker infrastructure, comprising mainly of proprietary lockers in commercial areas by a few large players. How do we design the LA and what proportion of the parcels in Singapore will be shifted to the LA? These are key questions to address before the nation-wide implementation.

Academic/Practical Relevance: Fixing other factors such as pricing and competition, we focus on the issues of network design and determine the appropriate coverage of the LA lockers. The problem is complicated by the fact that customers can pick up parcels either from the lockers near their homes or their work places. Without knowing the home-work office information of customers in the population, how do we determine the coverage of the LA network to ensure that its utilization will be maximized?

Methodology: We use a set of locker usage data from a proprietary system to calibrate a locker choice model, and use the observed delivery profile (without LA) to develop a facility location model for the network design problem. To solve the large-scale facility location problem, we propose an exact SOCP-MIP reformulation to this problem.

Results: We show that the volume of parcels delivered to the lockers, obtained without considering the issue of demand endogeneity, is a lower bound to the case when we factor in demand endogeneity. Contrary to conventional wisdom, our model may not always place lockers near areas with peak parcel volume because of demand endogeneity. As an application, we also provide an estimate to how much of the parcel traffic into the central business district (CBD) in Singapore can be diverted to LA lockers installed in housing estates outside of the CBD.

Managerial Implications: Using real data provided by the government agency and industry players, we show that with a well-chosen LA Network that satisfies the coverage requirement of the government, the volume of parcels delivered into CBD can be reduced by at least 7.5%. This will in turn contribute to a more sustainable city logistics ecosystem.

Keywords: Smart Nation; Last Mile Innovation; Facility Location; Demand Endogeneity

Suggested Citation

Lyu, Guodong and Teo, Chung-Piaw, Last Mile Innovation: The Case of the Locker Alliance Network (October 17, 2019). Available at SSRN: or

Guodong Lyu (Contact Author)

National University of Singapore (NUS) - NUS Business School ( email )

15 Kent Ridge Drive
Singapore, 119245

Chung-Piaw Teo

NUS Business School - Department of Decision Sciences ( email )

15 Kent Ridge Drive
Mochtar Riady Building, BIZ 1 8-69

Here is the Coronavirus
related research on SSRN

Paper statistics

Abstract Views
PlumX Metrics