Operations Research 2021
Abstract Submission

255. Vehicle routing for a population-wide COVID-19 testing program

Invited abstract in session TA-10: Logistics in the pandemic crisis, stream Logistics and Freight Transportation.

Thursday, 9:00-10:20
Room: Schreckhorn

Authors (first author is the speaker)

1. Emilio Jose Alarcon Ortega
Business Decisions and Analytics, University of Vienna
2. Margaretha Gansterer
University of Klagenfurt
3. Karl Doerner
Department of Business Decisions and Analytics, University of Vienna


During the recent pandemic scenario, a wide range of logistics problems arise. One of the recent strategies applied by national governments is related to gargling tests. Due to pooling techniques, these can be used to perform regular mass testings. However, in order to provide testing results in short time, an efficient logistics network has to be developed. We investigate such a network for a population-wide testing program initiated by the City of Vienna. In this program, all citizens are encouraged to perform gargling tests at home. Used testkits can be returned in retail stores and gas stations, from where they are picked up and delivered to a laboratory. We present a mathematical formulation that resembles the real-world problem related to the pick up of CoVid-19 gargle tests from these return stations, which are spread all over the City of Vienna. The problem is formulated as a multi-period vehicle routing problem where the customers require consistency in the arrival times at the locations and consistency of the drivers. Furthermore, it is possible to reduce the service times at the locations by increasing the number of drivers on each vehicle from one to two drivers. Furthermore, we study the impact of considering a transshipment location, as the laboratory is located outside of the city area. By introducing this transshipment node, we can reduce total costs. However, the transshipment point can store a maximum number of testkits that, afterwards, will be sent to the laboratory in a big capacity vehicle. To solve this problem, we propose a heuristic method that combines a cheapest insertion constructive heuristic with an adaptive large neighborhood search. An extensive computational study reveals interesting insights on this challenging real-world problem.


Status: accepted

Back to the list of papers